A Review of the Sources, Monitoring, Detection, and Removal of Typical Olfactory Substances Geosmin and 2-Methylisoborneol

Abstract

As the key sensory indicators for drinking water quality evaluation, the odor problems caused by geosmin (GSM) and dimethylisobornyl alcohol (2-MIB) have led to several major water supply crises around the world. In this paper, the theoretical framework of the whole process control of olfactory substances is systematically constructed through innovative research from multiple perspectives, and the main contributions are as follows: a comprehensive analysis of the sources of GSM and 2-MIB; an innovative summary of the monitoring methods of cyanobacteria and elaboration on the ways of controlling cyanobacteria in the water source; a comprehensive combing of the methods of olfactory substance detection technology, mainly the application of the new sensor technology; an in-depth summary of the techniques of olfactory removal; an analysis of the problems of the traditional water treatment technology; an analysis of the development and application of the new sensor technology, analyzing the advantages and disadvantages of traditional water treatment technology and classifying and elaborating on advanced oxidation processes (AOPs); and suggestions for the future research direction of each segment.

1. Introduction

The typical odor problem of drinking water cannot be ignored in current water treatment because the odor-causing substances geosmin (GSM) and 2-methylisoborneol (2-MIB) can be detected by humans at levels of a few tens of nanograms or a few nanograms. Although this peculiar-smelling water is not very harmful to the human body from a toxicological point of view, consumers will instinctively refuse to drink it. They no longer trust the quality of tap water and instead use bottled water or home water purification systems, which adds some economic burden. The economic burden caused by consumer avoidance behavior and additional treatment of drinking water odors totals CNY 290,690 ± 27,427 per million people per day, of which about 13% comes from the herd effect of people who are not sensitive to odors [1].

It is important for water utilities to provide the public with aesthetically acceptable drinking water. This is because consumers initially judge the quality of tap water based on its color, taste, and odor (T&O). Without training, they are good at spotting differences but cannot describe specific changes in water quality, taste, and odor [2,3]. The emergence of T&O problems not only affects the revenue of water companies, but also causes consumers to lose confidence in water quality and creates a psychological fear of water safety. It may even bring crisis to the entire drinking water industry and reduce the credibility of the government, thus becoming a problem of concern and attention around the world.

There are hundreds of odors in various water sources, which can be divided into three main categories and thirteen types. These include eight olfactory odors, four gustatory odors, and one oro-nasal odor. The four main types of odor are sour, sweet, bitter, and salty. Olfactory odors include earthy, musty, grassy, fishy, swampy, rotting, fruity, medicinal, and chemical tastes, as shown in Figure 1.

T&O compounds in water can be caused by various factors. The main factors are algal blooms and microbial metabolism and biodegradation. Two representative T&O compounds produced by algal blooms in surface water sources are 2-MIB and GSM, which give water bodies an earthy odor and are the most common cause of odor problems in drinking water obtained from natural waters [4]. In addition to thiol sulphides [5], β-cyclocitral and β-violet ketones [6], pyrazines [7], 2,4,6-trichloroanisole (TCA) and 2,4,6-tribromobenzyl methyl ether (TBA) [8], phenols and indoles [9], residual chlorine and some by-products from the chlorine disinfection process can also cause unpleasant odors [10], which is more obvious in some economically disadvantaged areas. Furthermore, the physical, chemical, and microbiological changes in drinking water during its transportation through the distribution network can also lead to the development of odors [11].

The most typical T&O event is triggered by 2-MIB, which produces musty odors. The odor threshold concentration (OTC) triggered by GSM is lower at 4 ng/L, while that of 2-MIB is 10 ng/L [12]. Their properties are shown in Table 1. In order to prevent the public from being exposed to the olfactory problems caused by GSM, countries have introduced policies to limit its concentration. Japan’s drinking water quality standard sets the acceptable concentration at 10 ng/L [13], while South Korea sets it below 20 ng/L [14]. China’s latest specification, “Standards for Drinking Water quality” (GB 5749-2022), also explicitly lists olfaction as a mandatory test item for ex-factory and piped water, and lists GSM and 2-MIB as extended indicators of water quality parameters, with a limit value of no more than 10 ng/L.

For GSM and 2-MIB, more research has been carried out on cyanobacteria as producers than on other sources such as actinomycetes. When an odor outbreak occurs, water utilities focus on removing GSM and 2-MIB from the water treatment perspective, while little attention is paid to monitoring and removing cyanobacteria at the source, and controlling cyanobacterial populations at the source can help water utilities take proactive measures before they receive a T&O complaint from a customer. Conventional testing techniques are limited by cost and operational complexity and are unable to perform rapid on-site testing for GSM and 2-MIB, and there is an urgent need to develop new testing instruments. In addition, conventional water treatment processes (CSFs) in water treatment plants do not have a better effect on the removal of GSM and 2-MIB, and advanced oxidation processes have been proven to be effective [15], but some of the advanced oxidation processes are costly and unsuitable for small- and medium-sized water plants. Therefore, in-depth studies on the sources, monitoring, detection, and removal of GSM and 2-MIB are urgently needed for a better response to odor incidents.

GSM and 2-MIB not only bring crisis to the drinking water industry, but also have a greater impact on the aquaculture industry. They are the most widespread odor metabolites in aquaculture, entering through the gills of fish and accumulating in the fatty tissues of fish, tainting aquatic products with an earthy odor. Although there is no documented evidence that they are harmful to humans through food, the T&O problems they bring can still cause substantial economic losses [16,17].

By understanding the sources, monitoring, detection, and removal of GSM and 2-MIB throughout the water supply chain, we can control GSM- and 2-MIB-induced odor events from multiple perspectives. This paper provides an in-depth review of GSM and 2-MIB, with the following aims: (1) To provide a comprehensive understanding of the production processes of GSM and 2-MIB through the study of the different producers of both. (2) To reduce GSM and 2-MIB at the source by monitoring the cyanobacterial population in water sources and by three removal pathways. (3) To summarize the new sensor technologies for fast and accurate on-site detection in addition to the traditional sensory and instrumental analysis methods. (4) To list the removal methods of olfactory substances by the traditional water treatment process and by a variety of advanced oxidation process techniques, with the aim of developing a suitable method for the removal of GSM and 2-MIB. We are committed to developing low-cost, green processes suitable for small- and medium-sized water plants.

2. Analysis of Typical Taste and Odor Incidents

The earliest research on typical odors in water originated from the Laurentian Great Lakes Basin in North America [18]. Odor events have occurred on every continent in the world, from Canada and the United States to Singapore, South Africa, Japan, the United Kingdom, and South Korea, as shown in Figure 2. Not only developed countries are plagued by this problem: developing countries are too. The rapid development of industry, the discharge of some wastewater, and improper management of water bodies are the main reasons for such outbreaks. A 2005 survey of 5 million Australia consumers involving 37 drinking water suppliers showed that around 78% of suppliers reported T&O problems with musty odors.

Along with the rapid development of China’s economy and the continuous progress of industry, many olfactory events continue to occur, the most typical of which are serious water quality events such as Taihu Lake, Dianchi Lake, Chaohu Lake, and the Luanjin Project. Chinese scholars’ extensive research on olfactory events began with the Taihu Lake water quality event in 2007, the four drinking water contamination events in Hangzhou in 2013, and the Lanzhou tap water oil spill odor event in 2014 [19].

Table 1. Typical odor incidents and major olfactory substances at home and abroad.

T&O Outbreak	T&O Burst Time	Typical Substances Causing Odor	Reference
Huangpu River, Shanghai, China	1990s	GSM, 2-MIB	[20]
Yangcheng Lake, Suzhou, China	1990s	2-MIB	[21]
Derwent River, Australia	2014–2017	GSM, 2-MIB	[22]
Mediterranean River Tel	For the past few years	GSM	[23]
Paldang Reservoir, Korea	Summer 2012	GSM	[24]
Bay of Quinte, Canada	-	GSM, 2-MIB	[25]
Taihu Lake, Wuxi, China	2007	Dimethyltrisulfide	[26]
Songhua River, Northeast China	For the past twenty years	-	[27]
Catalonia, Spain	2015	3-(trifluoromethyl)phenol	[28]
Fortaleza, Brazil Lake Shinji, Japan	2017 2007	GSM, 2-MIB GSM	[29] [30]
China’s Luan River flows into Tianjin	2015	GSM	[31]
Barcelona, Spain Diamond Valley Lake, USA Wichita Falls, USA	2002 2000–2004 2020	Diacetyl GSM, 2-MIB GSM	[32] [33] [34]

Table 1. Typical odor incidents and major olfactory substances at home and abroad.

T&O Outbreak	T&O Burst Time	Typical Substances Causing Odor	Reference
Huangpu River, Shanghai, China	1990s	GSM, 2-MIB	[20]
Yangcheng Lake, Suzhou, China	1990s	2-MIB	[21]
Derwent River, Australia	2014–2017	GSM, 2-MIB	[22]
Mediterranean River Tel	For the past few years	GSM	[23]
Paldang Reservoir, Korea	Summer 2012	GSM	[24]
Bay of Quinte, Canada	-	GSM, 2-MIB	[25]
Taihu Lake, Wuxi, China	2007	Dimethyltrisulfide	[26]
Songhua River, Northeast China	For the past twenty years	-	[27]
Catalonia, Spain	2015	3-(trifluoromethyl)phenol	[28]
Fortaleza, Brazil Lake Shinji, Japan	2017 2007	GSM, 2-MIB GSM	[29] [30]
China’s Luan River flows into Tianjin	2015	GSM	[31]
Barcelona, Spain Diamond Valley Lake, USA Wichita Falls, USA	2002 2000–2004 2020	Diacetyl GSM, 2-MIB GSM	[32] [33] [34]

Table 1 shows the more typical olfactory events, their outbreak time, and the substances that caused olfactory odor.

GSM and 2-MIB are the sources of these T&O events. A comprehensive survey of 111 drinking water treatment plants (DWTPs) in major cities in China was conducted, of which 80% of source water samples exhibited T&O problems, mainly characterized by soil/mold (41%) and swamp/septic tanks (36%), and the incidence of finished water was lower (45%). The source odor exhibited by river-based water sources was predominantly swamp/septic tank odor, whereas the source odor exhibited by lake- and reservoir-based water sources was soil/mold odor [18].

DCIP (bis(2-chloro-1-methylethyl) ether), 2E4MDL (2-ethy-4-methyl-1,3-dioxolane), and TMD (1,6-diamino-2,2,4(2,4,4)-trimethylhexane), as well as sulphuric ethers, cause the swampy/septic odor of river source water. The main causes of musty odor in lakes and reservoirs are GSM and 2-MIB. More than half of China’s water treatment plants that use lakes and reservoirs as water intake sites have olfactory odor problems [5,35,36]. Therefore, research on GSM and 2-MIB is urgently needed.

3. Sources of GSM and 2-MIB

GSM and 2-MIB are produced during the growth and metabolic secretion processes of cyanobacteria and some actinomycetes, and are also produced in small quantities by green algae, diatoms, slime molds, and fungi [37], as shown in Figure 3.

Cyanobacteria are some of the oldest organisms on Earth and are large unicellular prokaryotic organisms with a long evolutionary history. They are capable of oxygen-producing photosynthesis, have a wide distribution range, and strong adaptability. They are mainly divided into the genera Microcystis, Cichlidium, Chromococcus, Tremella, and Candida. Eutrophication of water bodies and climate change due to global warming, human activities, and nutrient loading from agriculture can lead to the formation of cyanobacterial blooms [38]. Cyanobacterial outbreaks can produce substances that affect water quality and cause odors, and produce toxins that are harmful to humans, such as microcystin-LR (MC-LR), columnaris toxin, staphylococcal clam toxin, and ichthyotoxin a [39,40]. Among them, microcystin is the most widely distributed and toxic toxin, with a lethal dose estimated to be 50 µg MC-LR per kilogram of body weight, and is the most commonly detected cyanobacterial toxin [41]. Currently, the World Health Organization (WHO) has set a maximum allowable concentration of total MC-LR in drinking water of 1.0 µg/L [42], which has prompted water suppliers to strive to develop effective toxin-removal treatments and management methods. MC-LR is highly water-soluble and therefore cannot be adequately adsorbed by activated charcoal or removed by other filtration steps and flocculation, so in actual operations deep treatment processes such as advanced oxidation processes are often used [43].

Cyanobacterial blooms tend to form in late summer (August), and in some cases as early as late spring [44]. Pseudanabaena sp. is the main producer of GSM and 2-MIB among cyanobacteria, with an intracellular 2-MIB content of up to 60% [45]. The dominance period of cyanobacteria is determined by the trophic state, light, and water temperatures of lakes or reservoirs [42]. Under adequate nutrient status, temperature significantly affects the growth of cyanobacteria and the synthesis of GSM and 2-MIB [46,47]. Cyanobacteria will become extinct at low temperatures, but when some low-temperature-tolerant algal blooms grow at the same time T&O events may occur, resulting in the problem of unsaturated aldehydes releasing flavor [48]. In lakes with inadequate trophic status, diatoms will be the more dominant algae.

During non-algal blooms, actinomycetes are usually considered to be responsible for olfactory events [49]. Actinomycetes are named as such because of their actinomycete-like colonies. They are widely distributed in nature and exist mainly in the form of spores or mycelia. They are important ecological decomposers in lakes and reservoirs, especially under hypoxic and anoxic conditions. They usually reach the peak in spring and autumn, and their abundance and distribution patterns are mainly affected by water temperature, and are also positively correlated with nitrogen and phosphorus [50,51,52]. Actinomycetes include Streptomyces spp., Nocardia spp., Actinobacteria spp., and Aeromonas spp. Among them, Streptomyces spp. are the main producers of GSM and 2-MIB [53], and Nocardia spp. can also produce GSM and 2-MIB [54].

Thermal stratification is a common phenomenon in lake and reservoir water sources, and an important regulator of phytoplankton communities. It directly controls the growth and succession of phytoplankton through temperature and nutrients, which in turn affects the concentration of olfactory substances [55]. Odors produced by phytoplankton tend to be near the water surface or in the mid-depth areas. For example, the new species of cyanobacterial bloom, Spirulina pumilus, mainly grows and secretes olfactory substances in the shallow waters of reservoirs [56]. In contrast, odors produced by actinomycetes are usually highest in concentrations near sediments. In addition, heavy rainfall is an important risk factor for increased odor in lakes and reservoirs [57], especially during periods of heavy rainfall following a dry period of the water table [53]. This may cause eutrophication of lakes and reservoirs, flushing large amounts of organic matter-rich material from land into lakes used for drinking water, increasing the likelihood of olfactory outbreaks [58].

4. GSM and 2-MIB Pre-Outbreak Monitoring

4.1. Monitoring of Algal Blooms

Outbreaks of olfactory substances in rivers and reservoirs not only affect the surrounding environment, but also the waterworks that draw water from them. For water utilities, establishing an excellent system to monitor the dynamics of potential producers in water sources is very helpful for early prediction of T&O outbreaks [59]. Monitoring of GSM and 2-MIB producers (mainly cyanobacteria) involves a wide range of technologies, ranging from traditional tools (such as microcounting, pigment extraction, quantitative real-time polymerase chain reactions (qPCRs), probes, and remote sensing) to emerging technologies (such as next-generation sequencing (NGS), metagenomics, photonics systems, biosensors, drones, and the application of machine learning) [60], as shown in Figure 4.

Previously, scholars used big data statistics to build mathematical models to predict the outbreaks of odorous compounds in the Great Lakes of North America, Taihu Lake in China, and Kansas State Reservoir in the United States [61,62,63]. A study [64] predicted the occurrence of algal blooms by constructing three-dimensional models and then predicted the production of odor compounds. Another study [65] combines machine learning to model and predict the outbreaks of cyanobacterial algal blooms.

Using advanced techniques to monitor algal blooms and odor compound outbreaks in water bodies is more economical and convenient than traditional field sampling. For example, real-time monitoring of algal conditions in water bodies through satellite remote sensing can quantify the chlorophyll a concentration simply by combining it with a calibration algorithm [66,67]. Machine learning technology is used to predict geosmin in reservoirs or lakes, where temperature and cyanobacterial abundance are important factors for predicting the surface geosmin concentration, and phytoplankton abundance is an important parameter for predicting the bottom geosmin concentration [68].

Sensor technology can be used for early monitoring due to its ease of operation and portability. Fluorescent sensors also be used to rapidly quantify cyanobacteria through probe readings. This method also relies on the measurement of chlorophyll a [69]. Chlorophyll a is the main photosynthetic pigment contained in all phototrophic microorganisms and may be measured inaccurately, which can be overcome by measuring specific cyanobacterial pigments. A study [70] used cyanobacteria’s proprietary pigment phycocyanin to measure the number of Pseudohyphae, which can be more reliable and accurate (up to 88% accuracy), and can therefore be used as an online early warning system for cyanobacteria in the intake area of reservoirs.

Many scholars have conducted research at the genetic level. A study [71] used molecular methods such as NGS and qPCR to investigate the diversity and dynamics of cyanobacteria. The NGS results of 2-MIB synthesis genes show that pseudoalgae and phytoplankton were the main producers of 2-MIB, with pseudoalgae being the dominant one. This finding suggests that NGS can quickly and accurately identify the producers of 2-MIB, thus playing an important role in the practical monitoring of aquatic ecology. Another study [72] conducted an in-depth field investigation of 2-MIB production events by Chlorella ashleyi in the Tsing Cho Sha Reservoir and found that the 2-MIB synthesis (mic) gene abundance (DNA) and expression (RNA) can be used as early warning parameters for 2-MIB production. The study also established an early warning model using mic gene abundance as an indicator of 2-MIB onset, and verified it in two other reservoirs, where light intensity is an important parameter influencing mic gene expression and MIB synthesis.

Key factors that must be considered in risk management monitoring of cyanobacterial blooms include changes in the diurnal vertical migration and community dynamics. Integrating multi-temporal point, multi-depth discrete sampling guides into lake and reservoir monitoring programs to characterize cyanobacterial community dynamics and signal changes can inform the risk management monitoring associated with cyanobacterial toxin production potentials, especially for nutrient-poor lakes. Cyanobacterial monitoring can also be enhanced by combining system characterization (e.g., thermal stratification) with community characterization [73]. In addition, the use of technology to monitor the amount of olfactory substances does not require complex pre-processing compared to analytical instruments, but the detection limit does not reach the level of precision instruments. Combining this with emerging technology to increase the detection limit, reduce costs, and make monitoring more accurate as well as convenient will be a major trend in future research.

4.2. For the Treatment of Cyanobacteria in Water Sources

By monitoring cyanobacterial outbreaks, there are methods that can be used to control cyanobacterial populations and address odor outbreaks at the source of the water treatment plant. Research shows that GSM and 2-MIB can be reduced at the source by controlling the production of cyanobacteria in the water column. Compared to removal using water treatment processes at the waterworks, it has better cost-effectiveness and changes the ecological status of water sources [74].

Cyanobacteria outbreaks can cause smells or lead to elevated odor levels. Controlling the number of cyanobacteria can solve the odor problem at the source. As shown in Figure 5, cyanobacteria can be controlled at three stages: at the water source of the water, in the pipelines transported to the plant, and after entering the water treatment plant.

Eutrophication of water sources is the main cause of cyanobacterial blooms. Long-term management goals can be achieved by reducing nutrient loading to the water body, limiting algal and bacterial growth, and thus reducing olfactory events [75]. A study [76] used solid sodium percarbonate (SPC) Na₂CO₃·1.5H₂O₂ as an additive to eliminate cyanobacteria distributed on the surface and bottom layers of the reservoir. It also used controlled dosages to achieve both good results and stability of the microeukaryotic community in the reservoir. The necessary factors before the bloom of surface cyanobacteria are high light intensity and nutrients [77], while for bottom cyanobacteria appropriately raising the reservoir water level is also an effective measure to reduce T&O events. Compared with the traditional physical, chemical, and biological techniques, this is a simpler and more economical management strategy. Moreover, it has no significant negative impacts on water quality or aquatic organisms [78]. Studies of minimum light growth requirements have shown that reducing the light availability for benthic/subterranean cyanobacteria may be an effective way to eliminate odor problems. Increasing the minimum light growth requirement by raising the water level or increasing the turbidity can limit benthic/subterranean cyanobacteria [79].

A study [80] analyzed 10 years of field data from the Oklahoma Drinking Water Utility and found that adding algaecide to the pipeline from the raw water to the water treatment plant can intuitively reduce the amount of GSM and chemically convert it into an odorless product. In addition, ultrasound can disrupt cell walls and cell membranes through cavitation, interrupt photosynthesis, and inhibit cell division and reproduction, and thereby inhibit the growth of cyanobacteria, which is much more environmentally friendly than using algaecides [81].

When cyanobacteria enter drinking water plants, oxidants such as ozone, potassium permanganate, and chlorine are used for pre-oxidation to improve the efficiency of subsequent conventional treatment processes (CSFs) by reducing the stability and activity of algal cells [82]. Some cyanobacterial toxins, especially MC-LR, can also be removed. High concentrations of potassium permanganate and chlorine can damage cyanobacterial cells, resulting in the destruction of intracellular organic matter, thus affecting the effectiveness of CSFs [83]. Cell rupture also causes GSM and 2-MIB to dissolve in a bound state. Contributing to the elevated concentrations of these two substances, cyanobacterial cells can enter the sludge because of CSFs, where they have been shown to survive for up to 10 days and produce large quantities of metabolites. Waterworks should also be aware of the significant risk that this sludge supernatant may pose when returned to the water [84]. Overgrowth of actinomycetes can cause sludge swelling in water plants [85], and the use of chlorine can solve the actinomycete problem by penetrating the cell membranes to inactivate the actinomycetes [49].

5. Detection and Analysis Methods

GSM and 2-MIB have a low olfactory threshold and produce effects at very low concentrations, so it is important to accurately detect and analyze their concentrations. Sensory analysis mainly uses humans as “sensors” to analyze the smell of drinking water in the early days, which is a qualitative analysis. In contrast, instrumental analysis can perform both qualitative and quantitative analysis. There are also new sensor technologies with low cost, fast control, and rapid analysis capabilities for on-site inspection.

There are three main organoleptic methods: flavor profile analysis (FPA), flavor rating analysis (FRA), and threshold of olfaction (TON). FPA classifies odor intensity into seven levels, with multiple surveyors analyzing and averaging the types and intensities of perceived odors according to an olfactory wheel diagram [86]. A study [87] used FPA to analyze the odor of water sources in several cities in the Yellow River Basin, and found that fishy, swampy, and moldy odors were present in these cities. The TON dilutes the water sample to an almost unnoticeable intensity, and uses a dilution multiplier to indicate the intensity of the odor. The FRA method involves treating a water sample, smelling the odor, and determining its intensity level according to six intensities and describing them in words. Sensory analysis is highly subjective, and prolonged sniffing can lead to olfactory fatigue and limitations [88].

Instrumental techniques for the qualitative and quantitative analysis of T&O compounds require high sensitivity and detection limits. Gas chromatography (GC) has a better detection on semi-volatile compounds. The flame ionization detector (FID) is not sufficiently sensitive to the concentration of T&O compounds, and is not selective enough for organic chemicals. The electron capture device (ECD) has selectivity and sensitivity but a narrow detection surface, and the mass spectrometry (MS) technique can take these problems well into account [89]. The integrated two-dimensional gas chromatography–time-of-flight mass spectrometry technique (GC× GC-TOFMS) and gas chromatography–triple quadrupole tandem mass spectrometry (GC-MS/MS) coupling technique are both commonly used for more accurate quantitative and qualitative analysis [90]. A study [91] developed an integrated two-dimensional gas chromatography with time-of-flight mass spectrometry based on the liquid–liquid extraction method for the Huangpu River and Huai River, and simultaneously identified 54 frequently encountered odor-prone compounds as sulphuric ethers, aldehydes, pyrazines, benzenes, and phenols. The method was shown to be able to simultaneously analyze different groups of compounds without derivatization and with much higher sensitivity.

Samples need to be pre-treated before they can be analyzed by the instrument, and they are pre-concentrated and extracted to improve the accuracy of the results. Commonly used pre-treatment methods include closed-loop stripping analysis (CLSA), open-loop stripping analysis (OLSA), purge tonicity (PT), liquid–liquid extraction (LLE), liquid-phase microextraction (LPME), simultaneous distillation extraction (SDE), solid-phase extraction (SPE), automated micro-solid-phase extraction (μSPE), solid-phase microextraction (SPME), stir bar sorptive extraction (SBSE), etc. [92,93,94].

The closed-loop stripping analysis (CLSA) and open-loop stripping analysis (OLSA) methods are simple, with a detection limit of up to 1 ng/L. The disadvantages of the methods are that the activated carbon needs to be replaced when measuring each sample, a large number of samples cannot be measured at the same time, and the change in the temperature during the operation will also affect the accuracy. Liquid–liquid extraction (LLE) and liquid–liquid microextraction (LLME) use a smaller volume of water samples, and have a lower cost and higher sensitivity, but the recovery and reproducibility of these methods are poor [95,96].

The purge tonicity (PT) method is green and environmentally friendly with high enrichment efficiency and no recontamination of organic solvents during sample processing. The time of purging is an important factor affecting the recovery and sensitivity of the method [97]. A study [98] synthesized novel materials as adsorbents for PT and combined it with GC-MS, analyzed five T&O compounds in water, and obtained good recoveries and detection limits. Solid-phase extraction (SPE) is a liquid chromatographic separation principle using selective adsorption and selective elution. Compared with liquid–liquid extraction, it is better at separating the analyte from its components, and is simple to operate, but takes a lot of time and water samples, and requires the use of large amounts of toxic organic solvents [99].

Solid-phase microextraction (SPME) can be used to extract target analytes from gaseous, liquid, or solid samples in smaller sample volumes and without solvents compared to LLE, SPE, and PT. Headspace solid-phase microextraction (HS-SPME) is currently the most commonly used technique for odor analysis. However, this method also suffers from the disadvantages of a more fragile fiber head, a limited number of coated adsorbents, and higher costs [100]. Stir bar sorptive extraction (SBSE) is the use of a closed glass magnetic stirring bar to stir the Dimethiconol(PDMS)-coated sample. The principle is similar to that of SPME, but with a higher sensitivity and a better linear range than SPME. [101]. Some pre-processing methods combined with detection techniques for GSM and 2-MIB are shown in

Table 2. Typical pre-processing techniques and detection methods for GSM and 2-MIB.

Analyte	Technology	Sample Volume	Extraction Phase	Recovery (%)	Detection Limit (ng/L)	Accuracy (%)	Reference
GSM, 2-MIB	μSPE- GC/MS	25 mL	C18	95.1–100.1%	GSM 2.0 2-MIB 4.3	<7%	[102]
GSM, 2-MIB, THMs	HS-SPME- GC/MS	5 mL	PDMS, CAR/PDMS, PDMS/DVB	80–120%	GSM, 2-MIB 5–50	<20%	[103]
GSM, 2-MIB, and 16 others	SPME-GC/MS/MS	10 mL	DVB/CAR/PDMS	70–120%	2-MIB 17 GSM 5	2–20%	[104]
GSM, 2-MIB, IBMP, IPMP	LLE, SPE	250 mL	Hexane and silica adsorption columns	84.6–103%	0.3–0.9	1.50–10.1%	[105]
GSM, 2-MIB	SPE-GC-MS/MS	1 L	IRIS PLUS	2-MIB > 90% GSM > 95%	0.9–5.5	GSM 8.5% 2-MIB 10.9%	[106]
GSM, 2-MIB, and 6 others	P&T+GC-MS	25 mL	Nitrogen at 40 mL/min	74.7–112.8%	MIB 0.3 GSM 0.2	2.6– 10.8%	[107]
GSM, 2-MIB, and 5 others	HS-SPME-GC-MS	40 mL	DVB/CAR/PDMS	2-MIB 84.1–119.0% GSM 87.7–95.9%	GSM 0.2 2-MIB 0.5	GSM 7.0% 2-MIB 4.9%	[108]
GSM, 2-MIB	SBSE-TD-GC-MS	-	PDMS	86–113%	0.2	<8%	[109]
GSM 2-MIB	SIDA-HS-SPME-GC/MS	-	PDMS/CAR/DVB	81–121%	GSM 3.0 2-MIB 3.1	GSM < 5.65% 2-MIB < 14.17%	[110]
GSM, 2-MIB, and 3 others	SPME-GC-ITDMS/MS	15 mL	DVB/CAR/PDMS	93–110%	<1.0	1–8%	[111]
GSM, 2-MIB, and 3 others	SPE-GC/MS	-	C18	93.5–108%	0.5–1.5	1.58–7.31%	[112]
GSM, 2-MIB, and 49 others	LLE-GC-MS	500 mL	-	70–120%	0.10–20.55	<20%	[113]
GSM, 2-MIB	USAD-LLME	12 mL	Tetrachloroethylene	70–113%	GSM 2.0 2-MIB 9.0	<11%	[114]
GSM, 2-MIB, and 6 others	PT-GC/MS	25 mL	Nitrogen at 40 mL/min	80.54–114.91%	<1.5	3.38– 8.59%	[115]

.

These pre-treatment and testing methods are expensive, time-consuming, require high levels of preservation of water samples, and are extremely inconvenient, making rapid or on-site testing impossible, and there is an urgent need to develop more convenient and less costly testing methods to cope with the practical demands of the job [116]. A study [117] achieved rapid detection of T&O compounds in water by measuring headspace vapor at room temperature using a newly invented chemical ionization coupled with TOF-MS, in which the limits of detection were 0.25 and 0.77 ng/L for GSM and 2-MIB, respectively, and which did not require a pre-enrichment process and had high sensitivity.

Sensory analysis is used for qualitative analysis, suitable for quickly determining the presence of odors in water and the immediate acceptance of drinking water by consumers, and instrumental analysis is used for quantitative analysis. Some scholars used FPA combined with SPE/SPME+GC-MS for the analysis of water in a reclaimed water treatment plant (RWTP) and found that the odor intensity was 6.4. Instrumental analysis confirmed that it was due to four compounds, namely dimethyldisulphide, dimethytrisulphide, indole, and 3-methylindole, with odor intensity (3.6) accounting for more than 50% of the influent odor intensity. The odor intensity of more than 50% after treatment was about 3.8, indicating that the water treatment process is more efficient in odor removal [118]. Water treatment plants are suitable for combining the two, which helps to better analyze water quality.

Sensor technology is used in several industries, mainly in the food industry as a means of distinguishing the freshness of food fruits, etc., and can be used to detect a variety of odorous substances including 2-MIB and GSM in the quality testing of domestic water and wastewater. The core principle is to convert these substances into measurable electrical or optical signals, etc. A study [119] investigated the detection of 13 typical odors such as β-violet ketone, β-cyclic citrulline, and other common odors by an electronic nose, among which 2-MIB and GSM were distinguished between 0 and 85 ng/L, and the response characteristics of the odors at different concentrations in pure and surface water were identified. The electronic tongue studied in the literature [120] can differentiate and quantify 10 T&O compounds including GSM and 2-MIB at concentrations as low as 0.02 µg/L. It can also detect algal toxins in water with high sensitivity to concentration changes and low cost. Compared with traditional unstable and complex detection methods, electronic noses and tongues can provide a promising alternative for early warning monitoring of odor events. A study [121] studied the binding properties and mechanisms of human odor-binding protein OPB2a with 14 typical odors. The results showed that OBP2a could bind nine odors, including IBMP and 2-MIB, and six aldehydes, providing a theoretical basis for the development of OBP-based odor-detecting biosensors to achieve rapid detection of drinking water olfactory odors.

6. Removal of Olfactory Substances from Drinking Water

The removal of olfactory substances has been a hot research topic in the field of water treatment, with the focus being on how to reduce the concentration to slightly below the human olfactory threshold. In natural water bodies, volatilization is a pathway to remove GSM and 2-MIB [122]. In addition, there are some bacteria in natural water bodies that can degrade these two substances, which is biodegradation [4]. These two degradation pathways do not lead to significant results. For water plants, the removal rate of the conventional treatment process (CSF) is less than 10%, while some traditional removal methods such as powdered activated carbon and biological methods are effective, based on which the combination process is constantly being studied and improved. In addition, there is the advanced oxidation process, which is currently popular because it can produce free radicals with strong oxidation ability. Through the strong oxidation ability there can be undifferentiated removal of these two olfactory substances, which can achieve better treatment results.

6.1. Traditional Methods of Removing Odors in Water Plants

Since GSM and 2-MIB are saturated cyclic tertiary alcohols with oxidation resistance, they are difficult to remove by the conventional treatment process, so chemical oxidants such as Cl₂, ClO₂, NaClO, and KMnO₄ have no effect on them. Powdered activated carbon, biological treatment, and ozone have been shown to be effective in treating GSM and 2-MIB. This section reviews the advantages and disadvantages of these three commonly used treatment technologies and their coupling processes.

6.1.1. Activated Carbon-Based Treatments

Large amounts of GSM and 2-MIB are produced during cyanobacterial outbreaks, with GSM mainly present in the cell-bound state and 2-MIB mainly dissolved outside the cells [4]. However, in water treatment plants the CSF is only effective against intracellularly bound 2-MIB, but has little effect on dissolved ones [123]. The addition of powdered activated carbon (PAC) is often used in water treatment plants to control the odor problem, which can be treated by controlling the dosage of PAC [124,125].

PAC can handle most odor problems and has the advantages of having a low cost being simple to use. However, it is not effective for odors caused by compounds such as dimethyl disulphide, diethyl disulphide, and dimethyl trisulphide disulphide [126]. In addition, organic matter produced by natural and algae metabolism will also compete for the adsorption sites of PAC, resulting in a decreased adsorption effect [127]. Furthermore, PAC is difficult to recycle, and the residual sludge after saturation is difficult to dispose of. The commonly used activated carbon thermal regeneration technology requires huge operating costs [128]. A study [129] investigated and tested three novel thermal reactivation procedures for activated carbon, namely steam curing, steam curing with elevated temperatures, and steam pyrolysis reactivation, for the removal of 2-MIB. These methods require lower temperatures and less replenishment of virgin charcoal than conventional activation methods.

Temperature also affects the adsorption effect of activated carbon, with a greater removal effect at 20 °C in summer than at 4 °C in winter [130]. Because GSM and 2-MIB are both hydrophobic compounds, the adsorption effect of low-hydrophilic activated carbon is better and is related to the particle size. Super-powdered activated carbon with a particle size less than 1 µm has a better effect, and the effect of any smaller size will be average. The effect of the hydrophilic activated carbon is not affected by the particle size [131,132]. GSM is more easily removed by activated carbon adsorption than 2-MIB. Activated carbon adsorption is more likely to remove MIB and GSM, but powdered activated carbon is only effective for dissolved GSM and 2-MIB, and not for intracellular ones.

When the concentration in the raw water is higher than a certain value, it is difficult to remove them below 10 ng/L by using powdered activated carbon alone, so combined processes must be adopted. Although the use of some pre-oxidants can increase the subsequent removal effect, it is also difficult to meet the water quality standard. A study [133] optimized the process of adsorption of PAC and pre-oxidation of potassium permanganate and found that the potassium permanganate remaining after pre-oxidation would oxidize the PAC, making the active site change and affecting the removal of 2-MIB. It is difficult to control the amount of 2-MIB below 10 ng/L with the pre-oxidation + PAC process, and thus some combination processes based on PAC have also received more attention. A study [134] synthesized a new photocatalyst of TiO₂-coated PAC (TiO₂/PAC) using the sol–gel method and found that the removal rate of 2-MIB under UV irradiation can reach 97.8%, which basically meets the requirements [135].

6.1.2. Biological Filtration in Biofilters with Biodegradable Geosmin and 2-MIB

GSM and 2-MIB do not easily degrade organic matter. Previous studies have shown that biodegradation can effectively solve this problem. In nature, some aquatic and soil microorganisms can use them as food and decompose them by metabolism [136]. In recirculating aquaculture systems, 2-MIB is mediated by both adsorption and biodegradation of sludge, with biodegradation regenerating adsorption sites as the primary process [137]. A study [138] found that three microorganisms, Rhodococcus erythropolis, Pseudomonas malodorans, and Rhodococcus xanthus, can effectively degrade 2-MIB. Other microorganisms from strains isolated from the surface of bioactivated carbon filters, such as Micrococcus, Flavobacterium, and Puccinia, have been shown to be effective degraders [139]. The main pathways of biodegradation are oxidation and dehydrogenation reactions [140].

GSM and 2-MIB can be removed by biofiltration, which does not require the addition of chemicals that can trigger by-products, and requires less technical means and infrastructure maintenance. However, it is highly dependent on the microbial species present. By inoculating and domesticating some microorganisms, these two substances can be degraded in bioactive sand filtration, and the degradation rate can be significantly increased [141,142]. There are also biofilters with GAC as the filter media, which has an adsorption effect in the early stage, and gradually converts to BAC as microorganisms attach. At this time, adsorption and biodegradation work together, and biofiltration plays a major role when the adsorption capacity is depleted. In biofiltration, empty-bed contact time (EBCT) is a significant variable for the effective removal of 2-MIB and earthy odors. When EBCT > 15 min, the removal of earthy odor is better. A high temperature and lower fluidized bed volume (BV) favors the removal of 2-MIB [143]. BAC has been shown to be more effective in removing various odors from aquatic plants, such as moldy, septic, fishy, grassy, and other odors [144], and the addition of pre-treatment sessions with oxidants such as ozone can significantly improve removal efficiency.

However, biological methods have a long treatment cycle, do not easily control pH and humidity, and cannot effectively deal with high concentrations of pollutants, so it is difficult to work effectively during olfactory outbreaks [145]. They are easily affected by environmental factors and have defects in their large-scale implementation, so further research is still needed. A study [146] isolated three bacteria with high degradation ability from the biofilter of a water treatment plant, and sequenced the whole genomes of the three isolates. Functional genes related to the degradation pathway were found in all of them, and their number was positively correlated with the degradation efficiency of 2-MIB. Research from the perspective of microbial genes will be an important research direction to overcome the various drawbacks of GSM and 2-MIB biological treatment.

6.1.3. Ozone Is a Common Removal Process in Water Treatment Plants

In addition to the traditional PAC and biofiltration that can effectively treat these two compounds, ozone can also achieve better results due to its strong oxidation ability [46]. Ozone reacts quickly with organic matter and has the effects of sterilizing, decolorizing and deodorizing, inactivating viruses, controlling algae, and effectively removing odors from water. However, ozone is not easy to store and transport, and when used in high doses it will produce bromate, which is a highly toxic by-product of brominated water disinfection. Typically, solid metal oxides can be used to catalyze ozone to increase its mineralization and reduce the formation of by-products [147]. It can also be combined with other processes during pre-treatment to achieve effluent water quality requirements [148].

Catalytic ozonation reactions have a higher removal efficiency and greater convenience compared to mono-ozone reactions [149]. Compared with ozonation alone or catalyst adsorption, raw mineral bauxite without further heat treatment shows no significant catalytic activity. In contrast, all further heat-treated bauxite catalysts exhibited some catalytic activity for the removal of 2-MIB. The change in crystalline during heat treatment plays an important role in improving ozone efficiency. The catalytic activity of bauxite roasted at 450 °C is the most significant, which consists mainly of γ-Al₂O₃. A study [150] investigated the mechanism of catalytic ozone oxidation of 2-MIB by alumina (γ-AlOOH and γ-Al₂O₃) and concluded that the role of surface hydroxyl groups in adsorption and catalytic ozone reactions determines the catalytic ozone reaction mechanism. The removal rate of γ-AlOOH and γ-Al₂O₃ for 2-MIB were 98.4% and 27.5%, respectively.

Ozone can also be used in combination with other water treatment processes. Compared with a single process, ozone oxidation + BAC can more effectively remove olfactory substances for both of these substances [151], but to remove odors from marshes, etc., additional subsequent treatment processes are required [152]. It can also be combined with membrane separation technology in the water treatment process. The use of ultrafiltration (UF) membranes alone is not effective in treating GSM and 2-MIB. After ozone pre-treatment, the combination of BAC + UF can almost completely remove the taste and odor. However, membrane separation technology also faces the problems of having a high cost and being easily contaminated [153]. A study [154] investigated ozone + biologically activated carbon (BAC) pre-treatment combined with UF membranes, which can significantly improve the removal efficiency of organic and odor substances.

The addition of hydrogen peroxide can solve the problem of bromate as a by-product of ozone oxidation. A study [155] installed a carbon-based cathode in the ozone primary solution and produced hydrogen peroxide at the cathode by incorporating electrochemistry, as shown in Equation (1). It reduces bromate formation and removes some naturally occurring organic matter, thereby improving the removal efficiency.

O₂ + 2H⁺ + 2e⁻ → H₂O₂

(1)

The removal of GSM and 2-MIB using ozone is also affected by pH. When the pH is high (>9), both GSM and 2-MIB are scavenged by the hydroxyl radicals produced by ozone. However, when the alkalinity of the water (bicarbonate and carbonate ions) is too high it acts on the hydroxyl radicals, reducing their number and thus reducing the efficiency of the reaction [156]. Ozone-based methods will emerge in future studies as combination processes and will better address cost and by-product issues.

6.2. Advanced Oxidation Process

Considering the poor removal effect of conventional chemical oxidants on GSM and 2-MIB, a study [157] investigated the use of ferrate (VI) for treating odor compounds brought by algae. The method is effective on olefinic T&O compounds and had no significant effect on non-olefinic GSM and 2-MIB. Many researchers have turned their attention to the advanced oxidation process (AOP), which has a stronger oxidizing power.

The advanced oxidation process (AOP) is receiving increasing attention due to its high treatment efficiency and potential to meet treatment demands. Its high treatment efficiency depends on the reactivity of free radicals, which is determined by the free genes. The free radical has at least one unpaired electron atom, molecule, or ion in its outermost electron shell, and thereby reacts with various highly active pollutants. This ability is expressed by the redox potential, where the standard redox potential E₀ is defined with respect to a standard reference hydrogen electrode, which is an arbitrarily given potential of 0v. The positive and negative values correspond to the oxidation and reduction abilities, respectively [158].

The free radicals produced by the advanced oxidation process have strong oxidizing power, such as hydroxyl radicals (OH E₀ = +2.8V), chloride radicals (Cl E₀ = +2.4V), sulphate radicals (SO₄⁻ E₀ = 3.1v), hydroperoxyl radicals (·HOO⁻), singlet oxygen (¹O₂), and superoxide radicals. The first three have been shown to have good removal capabilities, but single-linear state oxygen and superoxide radicals are somewhat deficient in removing GSM and 2-MIB. These radicals, through their strong oxidative capacity, indiscriminately degrade pollutants and mineralize them into non-polluting small molecules. Some scholars have applied the advanced oxidation process to remove olfactory substances in water and achieved relatively good results. Research on the advanced oxidation process for GSM and 2-MIB is also in progress, among which the advanced oxidation process that generates the first three radicals is more common.

6.2.1. Hydroxyl Radicals

In the conventional removal process mentioned above, ozone is the main hydroxyl radical generation to remove GSM and 2-MIB. In addition, hydrogen peroxide-based Fenton, electrochemical methods, semiconductor catalysis, nanobubbles, plasma discharge, etc., are essentially hydroxyl radicals generating and functioning.

Fenton

The Fenton method is widely used in wastewater treatment. Advanced oxidation processes such as electric Fenton, light Fenton, and thermal Fenton are derived from the traditional Fenton process. The Fenton method works better under acidic pH conditions. It classically uses Fe²⁺ as a catalyst to decompose H₂O₂, which in turn produces highly reactive ·OH. The reaction formula is shown in Equation (2):

Fe²⁺ + H₂O₂ → Fe³⁺ + OH⁻ + ·OH

(2)

Combination processes such as ozone + H₂O₂ [82] are effective in removing GSM and 2-MIB, and the addition of hydrogen peroxide can easily enhance the production of hydroxyl radicals [159], as shown in Equation (3).

2HO₂⁻ + 2O₃ + H₂O → 2OH⁻ + 3O₂ + HO₂· + ·OH

(3)

The removal effect of UV irradiation alone is poor, while vacuum ultraviolet (UV/VUV) has a better removal effect [160]. As shown in Equation (4), the removal rate of GSM and 2-MIB can reach more than 90% when combining UV hydrogen peroxide.

H₂O₂ + hv → 2·OH

(4)

A study [161] found that combining UV with the photo-Fenton method will enhance the removal effect. This is shown in Equations (2) and (5).

Fe³⁺ + H₂O₂ + hv → Fe²⁺ + ·OH + H⁺

(5)

Although this method has a better removal effect and can remove THMs and MC-LR, it requires a pH of 3. A study [162] studied the removal of GSM and 2-MIB by the UV/H₂O₂+BAC process. Both could be reduced to 5 ng/L, which is lower than the human olfactory threshold. The study also proposed the degradation pathways of both of them, which easily undergo the ring-opening, dehydrating, and demethylation reactions for GSM, while 2-MIB easily undergoes the dehydrating reaction, compared with the four degradation pathways that have been reported as dehydrating, C-C bond-breaking, demethylating, and coupling. The degradation reaction pathways and some of the intermediate products are shown in Figure 6 and Figure 7.

The preparation of hydrogen peroxide is complicated, and there are safety risks in terms of stability, transport, and storage, and it may decay during long-term storage [163]. To overcome these disadvantages, some scholars designed an electrochemical reactor for on-site generation of pure hydrogen peroxide and combined it with UV irradiation for the degradation of trace T&O compounds in drinking water, which has achieved good results [164]. The excess hydrogen peroxide when using UV/H₂O₂ can be removed by adding a BAC process [162].

Electrochemical Methods

The electrochemical method can generate the strong oxidizing active radical ·OH under the action of an electric field, thus increasing its effect of removing hard-to-degrade substances. Improving electrode materials can enhance its effect. According to the degradation mechanism, electrochemical oxidation technology can be divided into direct oxidation and indirect oxidation. The former refers to the pollutant adsorption to the anode surface, with the anode as the electron acceptor, and being oxidized by direct electron transfer. The latter is achieved through the generation of strong oxidizing substances at the anode, i.e., a variety of free radicals involved in the reaction.

H₂O → ·OH + H⁺ + e⁻

(6)

OH⁻ → ·OH + e⁻

(7)

A study [165] used Ti/IrO₂-Pt as the anode and 3.0 g/L NaCl as the supporting electrolyte at a current density of 40 mA/cm² to electrochemically degrade geosmin. After 60 min of electrolysis, the concentration of geosmin reduced from about 600 ng/L to 8 ng/L. In addition, under different initial pH conditions the electrochemical oxidation capacity of the Ti/IrO₂-Pt anode was maintained for a long time, which is a fast and effective method for treating olfactory substances. Hydroxyl radicals play a major role, and hypochlorous acid generated in the reaction promotes the degradation of deodorant. A study [166] developed electrochemical oxidation (EO) using boron-doped diamond (BDD) electrodes for the degradation of GSM and 2-MIB in a sulphate electrolyte, and sulphate solution is considered to be a good support electrolyte due to its low cost, high efficiency, easy accessibility, and not-easily halogenated properties. EO does not require more oxidants compared to other AOPs and is more environmentally friendly. Sulfate ions will generate persulfate ions in situ, and both the anion and cation levels are involved in the reaction, generating reactive oxygen such as hydroxyl radicals to participate in the reaction.

Semiconductor Photocatalysis and Electrocatalysis

Due to the good photochemical properties and lively electrochemical behavior of certain semiconductor materials, hydroxyl radicals will be generated to participate in the reaction after light and electricity catalysis. TiO₂ photocatalysis is a common advanced oxidation reaction using semiconductors. After TiO₂ is irradiated by a certain wavelength of light, the whole photocatalytic reaction can be decomposed into two half-reactions, i.e., the electron-induced reduction reaction and hole-induced oxidation reaction (h⁺+e⁻). The photogenerated electron e⁻ is easily captured by oxidizing substances such as dissolved oxygen in water to generate superoxide radical ·O₂⁻, while the photogenerated hole h⁺ can oxidize organic matter adsorbed on the surface of TiO₂ or first oxidize OH⁻ or water molecules adsorbed on the surface of TiO₂ into hydroxyl radicals. Then, the oxidative degradation of pollutants in water can be realized, and cause the mineralization of organic matter to form carbon dioxide and water [167,168], as expressed in Equations (8)–(11).

TiO₂ + hv → h⁺ + e⁻

(8)

h⁺ + OH⁻ → ·OH

(9)

h⁺ + H₂O → · OH + H⁺

(10)

e⁻ +O₂ → ·O₂⁻

(11)

TiO₂ in the suspended state is a better remover of olfactory substances than the stationary state. It can quickly photocatalytically remove GSM and 2-MIB in a suspended state, with a removal rate of more than 99% within 60 min [169]. A study [170] developed a pilot filled-bed photocatalytic reactor and evaluated the water treatment of laboratory and naturally contaminated samples. More than 90% of GSM and 2-MIB was removed under the human olfactory threshold, which has a high potential for application in the fisheries sector. A study [171] investigated the rate-affecting factors in TiO₂ photocatalysis removal of microcystin and geosmin, in which the decisive step was the production of ·OH in the valence band, rather than the redox in the conductivity band.

Electrochemical catalysis of semiconductors is to use these materials as electrodes to generate hydroxyl radicals at the anode to participate in the reaction. In addition to the ultraviolet light irradiation, a “hole” effect will be generated inside the TiO₂ semiconductor under the action of an external electric field. This method of combining photocatalysis with the photoelectric method to generate ·OH is also known as photoelectrocatalysis, which improves the activation efficiency compared with a single catalytic method [172]. Simultaneous doping of two or three dopants in titanium dioxide can significantly improve its photocatalytic activity. A study [173] investigated the effect of wavelength on the photocatalytic oxidation pathways of 2-MIB and GSM in the presence of Fe-N co-doped titanium dioxide, and a higher steady-state concentration of ·OH obtained at shorter wavelengths resulted in a higher T&O removal efficiency. Dehydration and ring opening are the main degradation pathways, with different photodegradation intermediates at different wavelengths. In most cases, for 2-MIB photocatalytic degradation intermediates with a ring structure are produced under visible light, while chain compounds tend to be produced under UV irradiation. GSM intermediates differ insignificantly, and these intermediates are not hazardous in the water column. In addition, the activation performance can be enhanced by coupling some other methods such as thermal activation [174].

Nanobubbles

Nanobubbles (NBs) technology is widely used in many fields such as environment, medical, energy, and chemical. Bubbles with a diameter less than 1 µm are called nanobubbles, which are classified into interfacial nanobubbles and bulk-phase nanobubbles. The latter are mainly used in advanced oxidation. NBs have an extremely small size, high internal pressure, high specific surface area, long stagnation time, and stability. A large number of free radicals (e.g., ·OH) are generated upon rupture, which removes pollutants from water. Commonly used preparation methods include mechanical shear, pressurized decompression, ultrasonic cavitation, and the turbulent tube method [175].

A study [109] used oxygen nanobubbles (NBs) in water to remove GSM and 2-MIB. Volatilization is the main mechanism for the removal of these two substances from distilled and deionized water. Oxidation of reactive oxygen can additionally remove them. When NBs are mixed with microbubbles (MBs), the formation of hydroxyl radicals is facilitated, which can remove GSM and 2-MBI by enhancing volatilization and oxidation at high temperatures, thereby increasing the removal of geosmin and 2-MIB.

Ultrasonic waves, shock waves, and jets can transiently cavitate water, causing it to decompose into ·OH and ·H. A study [176] used ultrasound-induced degradation of GSM and 2-MIB, which can rapidly and effectively degrade toxins produced by blue-green algae. Moreover, irradiation at 640 KH also effectively degraded GSM and 2-MIB. The analysis of the cleavage products suggested that local high-temperature pyrolysis and free radical pathways work together.

Nanobubble technology is basically chemical-free, with high oxidant utilization, few secondary pollutants, and good treatment of pollutants. However, its cost is higher than existing technologies and some other factors limit its application [177]. In addition, NBs also have some activating effects on PMS, and by adding NBs to the system of Cl⁻-activated PMS, NBs can effectively synergize with the Cl⁻/PMS system for the degradation of ionizable pollutants. The synergistic mechanism of NBs may be due to the fact that negatively charged NBs act as a “bridge”, which accelerates the reaction of Cl⁻-activated PMS and increases the contact between the active substances and pollutants, thus promoting the removal of pollutants [178]. NBs also have a broader application prospect in the removal of olfactory odor from water.

Plasma Technology

Plasma technology is an effective means in water treatment, and is divided into high-temperature plasma and low-temperature plasma. Low-temperature plasma technology remove pollutants from water through the high-pressure discharge generated by hydroxyl radicals and other reactive free radicals, as well as high-energy electrons, shock waves, and thermal decomposition. The main discharge methods of the low-temperature plasma method include corona discharge, dielectric blocking discharge, sliding arc discharge, etc. This method has a strong oxidizing capacity, high degradation rate, and fast reaction speed, which can not only remove typical odor-causing substances in drinking water, but also has a better removal effect for microcystins.

A study [179] used an atmospheric pressure underwater plasma system to remove GSM and 2-MIB as well as cyanobacteria. A porous hydrophobic ceramic tube was used to transfer a large number of generated plasma bubbles and short-lived actives through the micropores on the tube to the water to participate in the reaction. It effectively decomposed not only geosmin and 2-MIB, but also the cyanobacterial algae that produce this taste and odor compound. In this case, 2-MIB breaks down more slowly than geosmin, and both degrade more rapidly in river water than in distilled water. Although the plasma process may be more competitive than other AOPs, the high cost of equipment and energy consumption are constraints to its application.

6.2.2. Chloride Radicals

The chloride radical (Cl) has a potential similar to hydroxyl radical. It also has a strong oxidizing ability, and is thus used by researchers to remove typical odor-causing substances from water, often in combination with UV technology. Compared to the popular UV/H₂O₂ AOP treatment process, UV/chlorine AOP has a higher radical production efficiency and lower power requirements [180]. UV/H₂O₂ also has to be considered for quenching the excess hydrogen peroxide that increases the operating costs. Since the chlorine dose can be optimally maintained and does not require quenching or replenishment, it can be used as a disinfectant if there is excess, which can greatly improve the shortcomings of hydrogen peroxide [181]. During the whole reaction process, photolysis of free chlorine generates a large number of reactive free radicals involved in the degradation reaction [182]. In the whole UV/chlorine reaction system, ·OH, ·Cl, and ·Cl₂⁻ are the reactive ions reacting with the organic pollutants, and the reactions that take place are shown in Equations (12)–(15) [183].

HOCl + hv → ·OH + ·Cl

(12)

OCl⁻ + hv → ·O⁻ + ·Cl

(13)

·O⁻ + H₂O → ·OH + OH⁻

(14)

·Cl + Cl⁻ → ·Cl₂⁻

(15)

A study [184] using a combination of UV/chlorine showed that the process was able to remove 90% of GSM and 2-MIB within 5 min, while chlorine or light alone removed less than 20%. Although the hydroxyl radicals play the main role in oxidation, the chloride radicals still play an important role. The efficiency of the reaction is best in a weakly acidic environment.

6.2.3. Sulfate Radical

Advanced oxidation processes based on the generation of sulphate radicals are also commonly used removal methods, where persulfate (PS), including peroxymonosulphate (PMS) and peroxydisulphate (PDS), are relatively stable oxidizing agents that are good for storage and transport for on-site applications. Peroxynitrite produces poor oxidative degradation of pollutants through autolysis in most cases. It needs to be activated using some method such as a catalyst or energy to enhance the removal of pollutants by generating more sulphate radicals with a strong oxidative capacity (·SO₄⁻). They are more effective than the hydroxyl radicals because of its elective oxidation for electron transfer reactions. In addition, sulfate radicals have a long half-life (30–40 μs) compared to hydroxyl radicals and have been the focus of research because of their diverse catalytic activation, stability, economy, and easy recycling of solid catalysts. The generation process of sulfate radicals can be divided into two categories: energy activation reaction and catalytic activation reaction. The former mainly includes thermal activation, ultraviolet activation, ultrasonic activation, microwave activation, and radiolysis. The latter mainly includes electrochemical activation, carbon material activation, transition metal activation, and alkali activation [185,186,187].

Ultraviolet Activation

The UV-activated system first breaks the O-O bond in persulfate by energy input and produces reactive radicals, and then the water molecules absorb the energy of UV light to produce electrons and transfer them to persulfate, as expressed in Equations (16)–(20). A study [188] removed β-cyclocitraldehyde by using UV/PS. Another study [189] removed TCA with earthy and moldy smells by using UV/PMS. Both works achieved satisfactory removal rates, in which sulphate and hydroxyl radicals had degradation effects, but sulphate radicals were the main factor. Sulphate radicals have advantageous free radicals under acidic and neutral conditions, and their activation energy is minimized under neutral pH conditions, a property that makes them advantageous in water treatment. A study [190] compared UV/PDS with UV/H₂O₂, UV/NH₂Cl, and UV/HOCl for the treatment of GSM and 2-MIB, all of which achieved high removal rates, with UV/PDS being the most effective. The intermediates produced by the different substrate ring openings were different, but the transformation products were similar, with dehydration reactions playing an important role in the GSM and 2-MIB degradation. UV/PS has a strong oxidation capacity without secondary pollution, but it is susceptible to water quality, especially color and turbidity. The investment and operating costs of the equipment also limit its large-scale application.

S₂O₈²⁻ + hv → 2·SO₄⁻

(16)

HSO₅⁻ + hv → ·SO₄⁻ + ·OH

(17)

H₂O + hv → ·H +·OH

(18)

S₂O₈²⁻+ ·H → ·SO₄⁻ + SO₄²⁻+ H⁺

(19)

HSO₅⁻ + ·H → ·SO₄⁻ + H₂O

(20)

Thermal Activation

The mechanism of thermal activation is to promote the breaking of O-O bonds in persulfate by high temperatures to produce ·SO₄⁻ and ·OH to degrade pollutants. This method is simple and efficient without additional compounds. A study [191] investigated the mechanism and kinetics of the degradation of 2-MIB using thermally activated PS and Fe(II)-activated PS (PS-Fe²⁺). High temperatures and high initial concentrations of PS increased the decomposition rate of 2-MIB in the thermally activated PS system. The degradation rate was highest at pH 5.0. The process is represented in Equations (21) and (22). However, thermal activation requires a large amount of energy consumption, which limits its application scale.

S₂O₈²⁻ + heat → 2·SO₄⁻

(21)

HSO₅⁻ + heat → ·SO₄⁻ + ·OH

(22)

Ultrasonic Activation

The principle of ultrasonic activation is to activate the internal rapid generation of tiny bubbles within a solution and rupture through ultrasound. The local high temperature and high pressure make O-O bonds break, produce ·SO₄⁻ and ·OH, and then remove the organic pollutants. Ultrasonic activation can effectively avoid the problems such as harsh reaction conditions and easy secondary pollution that exist in other activation methods. A study [192] used ultrasonic activation of persulfate to remove carbamazepine (CBZ) from the aqueous solution. Ultrasonic irradiation and sulfate radical oxidation had a significant synergistic effect on the removal of CBZ, with the removal efficiency as high as 89.4% after 120 min of reaction. Sulfate radical was considered as the main oxidizing agent for the removal of CBZ in this system. The reaction formula is expressed in Equations (23)–(26). Enhanced ultrasonic power can increase the number of radicals for higher removal by increasing the number of cavitation bubbles and rapid activation of PMS. However, the main problem of ultrasonic activation is the high energy consumption, which is an important reason for limiting its large-scale application.

S₂O₈²⁻ + ))) → 2·SO₄⁻

(23)

H₂O + ))) → ·OH + ·H

(24)

HSO₅⁻ + ))) → ·SO₄⁻ + ·OH

(25)

S₂O₈²⁻ + ·H → ·SO₄⁻ + SO₄²⁻ + H⁺

(26)

Electrochemical Activation

The electrochemical activation of the persulfate electrochemical reaction generates sulfate radicals at the cathode by the same mechanism as the single-electron transfer redox reaction of iron-activated persulfate. The electrochemical activation of the PS process is mainly influenced by PS concentration, pH, current density, voltage, electrode material, and media [193]. The combination of electrochemical and electrochemical activation is superior to electrochemical techniques alone in terms of peroxynitrite removal, which can reduce the use of peroxynitrite and electricity, thereby reducing energy consumption. Electrochemical activation is green and efficient, but the electrodes are fragile, consume more energy, and are prone to produce hazardous substances, as expressed in Equations (27) and (28).

S₂O₈²⁻ + e⁻ → ·SO₄⁻ + SO₄²⁻

(27)

SO₄²⁻ → ·SO₄⁻ + e⁻

(28)

Transition Metal Activation

The persulfate activation by transition metals and their oxides has been studied for the removal of some special odors in water, such as Co²⁺, Fe²⁺, Mn²⁺, Ag⁺, Ni⁺, etc. Among them, silver ions showed the best activation effect on PDS, and divalent Co ions showed the best activation effect on PMS [194]. However, the use of heavy metal ions is prone to secondary pollution and they are difficult to be recycle. Iron, the most widely studied metal, is an effective activator, relatively non-toxic, environmentally friendly, and more cost-effective than other transition metals, and thus Fe²⁺ is generally chosen to activate PS. A study [195] conducted a comparison of the activation of PMS and PS with Fe²⁺ to remove two odor compounds, β-cyclocitraldehyde and TCA, which give fruity and musty tastes to drinking water. The odor compounds can be removed well compared to coagulation treatment with Fe(II) and Fe(III) alone, where PMS is more effective than PS. Increasing the concentration of PMS and Fe(II) can effectively remove NOM and control the production of DBPs. The equations for transition metal-activated persulfate are shown in Equations (29) and (30):

Mⁿ⁺ +S₂O₈²⁻ → M ⁽ⁿ⁺¹⁾⁺ + ·SO₄⁻ + SO₄²⁻

(29)

Mⁿ⁺ +HSO₅⁻ → M ⁽ⁿ⁺¹⁾⁺ + ·SO₄⁻+ OH⁻

(30)

In addition to monometallic activation of persulfate, composite metal catalysis has more active sites than monometallic, which is beneficial to improve the activation efficiency [196]. The second metal is usually less active and is thought to promote iron oxidation and may act as a catalyst for electron transfer and hydrogenation. Zero-valent iron can also activate persulfate. It can be used as a source of ferrous iron in the reaction system, and can also convert ferrous ions in the generated product into ferrous ions for recycling. In practice, zero-valent iron with a diameter of up to the nanometer scale will have a better activation effect due to its large specific surface area [197]. The reaction equation is shown in Equations (31)–(34).

2Fe⁰ + O₂ + 2H₂O → 2Fe²⁺ + 4OH⁻

(31)

Fe⁰ + 2S₂O₈²⁻ → Fe²⁺ + 2·SO₄⁻ + 2SO₄²⁻

(32)

Fe²⁺ + S₂O₈²⁻ → Fe³⁺ +·SO₄⁻ + SO₄²⁻

(33)

Fe⁰ + 2Fe³⁺ → 3Fe²⁺

(34)

Carbon Material Activation

Carbon materials (graphene, carbon nanotubes, activated carbon, biochar, etc.) have the advantages of being metal-free, having a high utilization, being acid- and alkali-resistant, being widely available, being green, having a low cost, having a large specific surface area, having abundant active sites, etc., which make them a better choice for activating persulfates. The carbon materials for activated persulfate are divided into three categories: original carbon materials and their derivatives, heteroatom-doped carbon materials, and metal particle carbon materials.

For pristine carbon materials, the presence of persistent free radicals, oxygen-containing functional groups, and sp2/sp3 hybridization defect structures on the surface of the material is the main activation mechanism. The single-linear state oxygen (¹O₂) generated from the activation process can also be the main active species involved in the degradation of pollutants by the non-radical pathway [198,199]. Modification of materials by doping carbon materials with heteroatoms (N, S, B, or P) can significantly increase the activation capacity. The introduction of heteroatoms can add more active sites, enhance electron transfer, and increase defect edges [200]. Loading metals can also increase the activation performance, especially when loading transition metals, which have activation effects themselves and will have synergistic effects with carbon materials, avoiding the phenomenon of secondary pollution caused by leaching of heavy metal ions [201]. A study [202] achieved good results in the tetracycline removal by activation of persulfate with biochar material loaded with FeOX. For charcoal materials and acid, alkali, and thermal modification treatments, modification is an important method of enhancing the activation effect.

Joint Activation

Some scholars have explored the use of persulfate as a substrate in advanced oxidation processes to remove emerging contaminants such as antibiotics and perfluorinated compounds [203]. However, in terms of water treatment the addition of persulfate during the treatment process will inevitably lead to higher levels of cations and sulfate ions in the effluent of the water plant. This may require additional subsequent treatment processes, increasing the corresponding costs of the water plant. Activation by energy inevitably requires a large amount of energy consumption, and metal activation involves the risk of leaching of metal ions. Therefore, it is a good direction to study how to enhance the activation effect of persulphate, reduce the dosage of persulphate while achieving the same effect, and avoid various unfavorable factors. Multi-factor combination is a better treatment method.

A study [204] investigated P-doped treated biochar loaded with nano zero-valent iron for the ultra-efficient activation of persulfate to remove pollutants. The small particle size and magnetic properties of nano zero-valent iron make them prone to agglomeration and the high reducing activity makes them susceptible to oxidative deactivation by airborne oxygen. Unmodified Fe-Zr nanoparticles have poor mobility in groundwater environments. Using biochar loading can reduce agglomeration, delay oxidation in the environment, and improve transport in the porous medium. In addition, ultrasound, heat, and Fe²⁺ were also used to activate persulfate simultaneously to remove SO₂ and NO from flue gas. The NO removal efficiency increased with increasing persulfate concentration, ultrasound power density, and Fe²⁺ concentration (high persulfate concentration). Ultrasound, Fe²⁺, and heat had a synergistic activation effect, which enhanced the activation efficiency compared to a single activation method [205].

Other scholars have also studied the biochar, bimetallic combined with electrochemical co-activation of persulfate. A study [206] using a bimetallic modified biochar (MBC) cathode (MBC@CF) as the main body, the removal of tetracycline (TC) was achieved through the electric field enhancement of catalytic cathodic activation of PMS (E-MBC@CF-PMS). The experimental results showed that the activation effect on PMS under this system could be improved by two-to-three times compared with single electric activation, and the combined use of multiple activation techniques has an obvious enhancement effect and is an indispensable means in advanced oxidation research.

6.3. Summary

Although higher concentrations of oxidising agents give better results, excessive amounts will cause a waste of resources. Moreover, each removal process has its own advantages and disadvantages. It is worth studying how to optimize to the appropriate dosage as well as develop a new type of low-cost and environmentally friendly removal process. For example, in the study of advanced oxidation based on persulfate, the selection of different electrode materials for electrochemical activation involves the cost problem. Rare metals may be more effective, but the cost may be too high. In the study of the modification and optimization of carbon materials for activation of PMS, the selection of different raw materials for carbon and different modification methods will affect the activation effect. Moreover, the raw materials of biochar are simple and easy to obtain, with a low cost, and the modification method is also relatively simple and has better application prospects.

In the past, the research on these two kinds of olfactory substances was carried out in the laboratory, but in the future practical application in water treatment plants should be the main focus. The actual situation in water treatment plants is more complicated. The odor comes not only from the raw water, but also from some organics converted into odor during the process treatment. Odor is also produced during the pipeline network transportation due to some other reasons [143]. In practical applications, the influence of natural organic matter is also inevitable. From the perspective of water treatment, natural organic matter is mainly humic and xanthic acids, which consume the free radicals generated by the advanced oxidation process, and greatly affect the treatment effect of the advanced oxidation process, while generating disinfection by-products that affect human health [207,208].

The current process for coping with water quality exceeding algal counts in full-process waterworks is mostly a dual-depth treatment process with ozone-activated carbon + ultrafiltration membrane treatment after pre-treatment with ozone. The conventional water plant treatment process (CSF) is followed by disinfection with UV; however, large-scale application of UV is still difficult due to the energy cost of UV supply and the lack of suitable equipment [140]. Therefore, full-flow process water plants are only concentrated in economically developed large cities, and for small- and medium-sized plants, how to deal with the GSM and 2-MIB outbreaks still needs to be improved and researched.

7. Conclusions and Recommendations

In summary, this study systematically sorted out the whole process control system of GSM and 2-MIB, elaborated the main producers of both cyanobacteria and other potential producers through traceability analysis, innovatively summarized many monitoring means and treatment methods for source cyanobacterial outbreaks, compared the applicable boundaries of three detection methods, namely, sensory analysis, instrumental analysis, and sensor technology, listed the traditional water treatment processes and emerging advanced oxidation processes, and constructed a multilevel prevention and control framework from source early warning monitoring and treatment to transportation to water treatment plants.

The monitoring component lies in the development of new biosensors in conjunction with new technologies and research at the genetic level. The monitoring should take into account various aspects such as the migration of cyanobacteria, dynamic changes, and other influencing factors such as the thermal stratification of lakes, so that the control of blooms at the source and the prevention of cyanobacterial blooms is an effective way of controlling the GSM and 2-MIB.

For the detection of the two odor-causing substances, the development of new low-cost sample pre-treatment methods based on solid-phase microextraction with a high recovery rate will be the research direction. Sensory and instrumental analyses can be complementary to each other, but are costly and time-consuming. It is recommended to develop new electrochemical-based sensors to improve the detection limit for rapid on-site detection.

Regarding the treatment process of GSM and 2-MIB, activated carbon adsorption and ultraviolet method are more green, and the advanced oxidation process based on hydroxyl radical and sulfate radical is the current mainstream, but some processes will have residues, photocatalysis, electrochemical activation, nanobubbles, and carbon material activation, which are more green activation methods and will be an important research direction. It is recommended that some green activation methods be coupled to improve the activation effect and reduce the residues. In addition, some new low-cost materials can be developed for membrane filtration. The development of new processes mainly for small- and medium-sized water treatment plants can be put into use.