Nitrogen and Phosphorus Removal from Wastewater Using Calcareous Waste Shells—A Systematic Literature Review

: Nitrogen and phosphorus in freshwaters are a global environmental challenge. Concur-rently, the shell ﬁ sh industry’s calcareous waste shells (CWSs) amount to ~10 million tonnes annually. CWSs can e ﬀ ectively adsorb dissolved pollutants, including nutrients, from water, which has motivated a growing number of experimental studies on recycling CWSs in wastewater treatment. This comprehensive literature review summarises and critically assesses the e ﬀ ectiveness of using di ﬀ erent CWSs for removing nutrients from water. The e ﬀ ects of CWS type, initial pollutant concentration, adsorbent dosage, particle size, and contact time (CT) are investigated. The results show that phosphorus removal has been examined more than nitrogen. Most studies have been conducted using synthetic wastewater under laboratory conditions only. There is a large variability in experimental conditions, such as CWS adsorbent dosages (0.1–100 g/L) and CT (0.083–360 h). The calcination of CWSs is frequently used to enhance adsorption capacity. The Langmuir isotherm model has been found to ﬁ t adsorption data best when raw oyster shells are used, while the Freundlich isotherm is best when the adsorbent is calcinated mussel shells. The pseudo-second-order (PSO) kinetics model tends to describe adsorption data be tt er than the pseudo-ﬁ rst-order (PFO) model in all shell types. There is signi ﬁ cant potential for using calcareous waste shells to remove nutrients from wastewater in line with circular economy aspirations.


Introduction
Nutrient pollution is the current primary cause of water quality impairment globally [1].Persistent and excessive nitrogen and phosphorus inputs to freshwater systems lead to eutrophication.This is typically caused by rapid urbanisation [2,3], discharges from under-performing wastewater treatment systems [4,5], and excess fertiliser runoff from agricultural intensification [6][7][8].As a result, excessive algal growth and dissolved oxygen depletion occur, causing an imbalance in aquatic ecosystems and diminishing freshwater resources [9,10].Concentrations of 0.1 mg/L of phosphorus [5,11] and 2 mg/L of nitrate [12] are sufficient to cause freshwater eutrophication.The USEPA especially recommends a chronic criterion concentration of 2.4 mg of NH4 + /L at pH 7 and 20 °C for a 30-day average duration, not to be exceeded more than once every three years on average [13].For comparison, the maximum allowable nitrate and ammonium concentrations in drinking water specified by the World Health Organization (WHO) are 40 mg/L and 0.2 mg/L, respectively [14].
As freshwater nutrient challenges are escalating, wastes from seafood processing industries are concurrently increasing.In 2020, the commercial shellfish industry was estimated at 42.6% of global seafood production by the UN's Food and Agriculture Organization (FAO) [15].Shellfish and crustacean aquatic products comprise 70% calcareous shells, with the remaining 30% comprising edible meat [15].In New Zealand alone (pop.~ 5 million), a Pacific Island nation, commercial fishing and aquaculture industries generated NZD 2 billion (0.7% of the country's GDP) in 2019 [16].Of this, approximately 100,000 tonnes/year of green lip mussels were processed for meat, with the shells becoming waste material requiring disposal [17].The New Zealand aquaculture sector is projected to grow from NZD 600 million in 2019 to NZD 3 billion by 2035, and other countries are similarly growing their shellfish production, so finding new ways of reusing and valorising CWSs will become critical to supporting more sustainable aquaculture [15,18,19] and aligning with aspirations to adhere to the sustainable development goals.
The use of CWSs to treat various wastewater pollutants has grown over the last few years.Morris et al. [33] and Nguyen et al. [18] treated commercial dyes and heavy metals, finding that Zn 2+ was more easily removed than Pb 2+ and Cd 2+ .Nutrient removal using CWSs has also been studied, with phosphorus removal attributed to adsorption and/or precipitation [4,6,34] and the removal of nitrogenous compounds, such as ammonium (NH4 + ) and nitrate (NO3 -), attributed to chemisorption [18,35,36] or physisorption [10].In most of the studies that trialled CWSs as an adsorbent [20,37,38], the calcination of CWSs at 500-1000 °C was applied to increase their reactivity [19,33].Depending on the CWS calcination temperature and the reaction product formed, phosphate is removed via adsorption on calcite [6], or by precipitation as calcium phosphate when phosphate ions react with calcium ions from the dissolution of calcium oxide in water [39].However, Nam et al. [40] found that hydrated calcinated oyster shells did not significantly remove nitrogen, due to the high solubility of nitrogen compounds compared to the near-complete phosphorus removal via alkaline precipitation at pH 12.
Bhatnagar and Sillanpää [41] reviewed nitrate removal from wastewater using different adsorbents but did not include CWS adsorbents, focussing more on activated carbons, clay-based adsorbents, double-layered hydroxides, zeolites, agricultural waste materials, and industrial wastes.Recent review papers have discussed valorisation opportunities for CWSs from snail shells [42] and molluscan shells in wastewater treatment [43], and other shell wastes, i.e., bivalve shells [44], egg shells [45], and general shellfish waste [15], in multiple applications in agriculture, construction, environmental protection, biomaterials, and food additives.Hart [21] also provided a mini-review on the properties and various applications of waste shells in wastewater treatment, cement production, building construction aggregates, biomedical manufacturing, and industrial catalysts.Eggshells have been used in a broad range of applications including in polymer, metal, and ceramic composites, as evaluated by Vandeginste [45].Wan Mahari et al. [46] reviewed studies that used CWS biochar as an adsorbent in removing emerging pollutants from aquaculture wastewater.The conversion of large amounts of widely available CWSs from the food industry, such as eggshells, into useful resources contributes to reducing, reusing, and recycling this waste and, thus, to the circular economy.Despite these reviews on CWS valorisation opportunities, there is no comprehensive review on the use of CWSs to remove nutrients from wastewaters.This comprehensive, systematic review was therefore undertaken to synthesise and critically assess the current knowledge and research gaps in using CWSs as an adsorbent to remove nitrogen and phosphorus from wastewater.The main objectives are to (i) synthesise the literature on nutrient removal from wastewaters by CWSs; (ii) summarise the nutrient removal mechanisms, isotherm, and kinetic models in the studies; (iii) examine the effects of operating parameters (wastewater type, adsorbent dosage, contact time, pH, pollutant concentration, and particle size) on nutrient removal; and (iv) highlight knowledge gaps and future directions for the successful use of CWSs as an adsorbent for water nutrient removal.

Materials and Methods
A comprehensive search of the scientific literature published from 2010 to 2022 was conducted using Scopus and Compendex databases.Scopus is an extensive, multidisciplinary, bibliographic database, and its use has been widely reported for systematic literature reviews [47][48][49][50][51]. Burnham [52] also stated that Scopus is preferred to the Web of Science (WoS) due to its higher indexed citations and daily database updating.Scopus provides a broader and more inclusive content coverage in the engineering, chemistry, environment, and general sciences, with easier open access to authors [52,53].
The screening of the literature included searching for key terms in the title, abstract, and keywords.Terms included calcareous shell, mussel shell, oyster shell, clam shell, eggshell, or bivalve shell (Step 1, Figure 1).Literature types were confined to journal articles and conference papers (Step 2, Figure 1) for the last decade (2010-2022) and published in English (Step 3, Figure 1).Keywords were then filtered to nitrogenous and phosphorus compounds (Step 4, Figure 1), including wastewater (Step 5, Figure 1).Next, the search criteria focussed on the treatment or removal of those pollutants using CWSs (Step 6, Figure 1).Finally, manual screening excluded articles about other pollutants such as heavy metals, bacteria, or soil (Steps 7 and 8, Figure 1).
After this screening process (Figure 1), three specific criteria were applied in Step 9 (Figure 1) to refine the search further: (i) Does the study investigate nutrient removal from wastewater using CWSs?(ii) Does the study report pollutant removal capacity? (iii) Does the study discuss the possible removal mechanisms, isotherm, and kinetics?Once all these three criteria were satisfied, the total number of publications analysed in detail was 64 (Step 10, Figure 1).Relevant information from these publications was then extracted into a custom-built Excel spreadsheet and grouped into different tables and figures for data analysis and interpretation.Further data analysis involved the use of OriginPro 2024b software for data presentation in a box plot and Pearson's correlation.

Regional and Temporal Trends
Most publications (53%, n = 34) originate from China and Malaysia (Figure 2a).This may be a reflection of the sheer volume of scientific publications emerging from China (in particular) with a large population and active research community [59].Additionally, it may be because these (and other cited countries such as Brazil, Japan, Malaysia, South Korea, Taiwan, and Vietnam) have long coastlines with booming marine farms, and hence have an interest in valorising abundant seashell waste [15,33,37,42,60].There has been a sharp increase in the number of studies using CWSs to treat nutrients since 2015 (Figure 2b), with this trend expected to continue because of the increased global awareness of freshwater pollution [19] and CWS waste valorisation opportunities [7,36].Furthermore, there has been a recent trend to investigate CWS biocomposites for the purpose of removing nutrients from wastewater, such as testing eggshell and potato peel [27], oyster shell and rice husk biochar [28], oyster shell and tobacco straw biochar [61], and eggshell and palm mesocarp fibres [62].[67] NH4 + , PO4 3− (lake) [68] PO4 3−  [69] Green-lipped mussels Perna canaliculus Green-lipped mussels Perna canaliculus Green mussel shells Perna viridis NH3-N (raw leachate) [30] Green mussel shells Perna viridis NH3-N (raw leachate) [70] Green mussels NH3-N (raw leachate) [

Wastewater Type and Form of Treated Nutrient
Calcareous waste shells have been investigated more for phosphorus removal (70% of studies) than for nitrogen (30%) (Figure 2c), possibly because phosphorus is easier to remove via adsorption, chemisorption, and precipitation compared to nitrogen.Phosphorus can be expressed as phosphate (PO4 3− ), phosphate-phosphorus (PO4-P), elemental phosphorus (P), and total phosphorus (TP), while nitrogen can be stated as ammonia (NH3 + ), ammonium (NH4 + ), ammoniacal nitrogen (NH3-N), nitrate (NO3 -), nitrate-nitrogen (NO3-N), and total nitrogen (TN) (Table 1).Most authors used synthetic wastewater (Figure 2c) in their experiments to remove nutrients (P-68%, N-26%), followed by domestic/municipal wastewater (P-20%, N-26%).A few reasons may explain the predominance of using laboratory synthetic wastewater in these studies.Firstly, real wastewater is inherently variable and chemically complex [42], so using synthetic wastewater, especially in early-stage laboratory experiments, is simpler and helps reduce confounding experimental effects.Secondly, access to and the potential pathogenic risks of using real wastewater can impede the progress of investigations due to health and safety concerns and logistical complexity.Lastly, field experiments would likely use actual wastewater, require larger volumes, and would be influenced by prevailing climatic variables, so they have more logistical requirements.Most of these studies used small wastewater volumes (between 15 and 2000 mL) (Table S1, Supplementary Materials) and were bench-top-scale.They focussed on the effects of different experimental parameters (e.g., dosage, HRT, and temperature) on nutrient removal [69,95] and identified the nutrient removal kinetics at play [76,93].Of all the non-synthetic wastewaters, landfill leachate was used to investigate nitrogen (especially ammoniacal nitrogen) removal, possibly because landfill leachate is characteristically high in ammonia [30,36].
The different nutrient forms investigated, as a function of the different CWS types, are detailed in Table 1 and summarised in Figure 2d.Phosphorus (predominantly phosphate) removal was investigated using all eight CWS types, including a combination of shells, whereas nitrogen removal was only tested by three CWS types: oysters, mussels, and eggshells (Figure 2d).It is likely that a greater number of studies investigated phosphate removal because it is well reported how phosphate is removed by adsorption or chemisorption under calcareous conditions [7,60,91].During adsorption, trivalent phosphate anions (PO4 3-) bond more strongly with divalent calcium cations (Ca 2+ ) compared to monovalent nitrate anions (NO3 -), resulting in faster and stronger phosphate removal [60].
The total nitrogen (TN) concentration was only measured in two studies (Figure 2d, Table 1), both using oyster shells in biological filter systems with aerobic nitrification in the upper layer(s) and anaerobic denitrification in the lower biofilter layers [79,84].Conventional biological aerated biofilters (BAFs) did not remove TN, due to the absence of an anoxic stage [84].However, raw oyster shells incorporated as alkaline filters were reported to maintain the anoxic system's neutral pH [22] and act as a biocarrier for denitrifiers [84].The removal of other nitrogenous (ammonia and nitrate) forms was investigated using oyster shells (12 studies), mussel shells (7 studies), and eggshells (4 studies) (Figure 2d).In the absence of a microorganism-driven system, the chemisorption of ammonia occurred heterogeneously on multilayers of the shells [18].In contrast, Daud et al. [72] and Detho et al. [70], who investigated ammonia removal from landfill leachate, deduced that a composite adsorbent containing mussel shells provided an increased surface area for greater monolayer adsorption compared to mussel shells only (167 mg/g).This composite adsorbent achieved a 60% reduction and adsorption capacity of 200 mg/g of ammoniacal nitrogen.Bhatnagar and Sillanpää [41] found that surface area does not substantially influence nitrate adsorption but that positively charged surfaces do, presumably as this increases the binding of nitrate anions.They found better nitrate adsorption with adsorbent protonation (activated carbon, zeolites, industrial wastes), ionic exchange (clay-based adsorbents, double-layered hydroxides), and an increase in the presence of amine and hydroxyl functional groups (chitosan, agricultural waste materials).
Most studies reported the removal of a single nutrient, and less emphasis has been placed on the simultaneous removal of nitrogenous compounds and phosphate.As mentioned earlier in this section, phosphate removal is well documented compared to that of nitrogenous compounds.Nevertheless, the successful removal of both nutrients from domestic or municipal wastewater using biofiltration [85,88] and coagulation/flocculation [60] was reported.In general, CWSs in biofilters remove phosphate via chemical sorption and precipitation, while the removal of nitrate and ammonium is facilitated by bacteria during nitrification and denitrification processes [85].On the other hand, when CWSs are used in coagulation/flocculation, its main constituent, calcium oxide (CaO), transforms into calcium hydroxide (CaOH) upon hydration, which simultaneously removes phosphate and nitrate via the precipitation of calcium phosphate (hydroxyapatite) and calcium nitrate-hydroxide complexes, respectively [60].A more detailed explanation of both nutrient removal mechanisms can be found in Section 3.5.

Effects of Experimental Conditions on Nutrient Removal
Table S1 (Supplementary Materials) summarises the key experimental conditions and nutrient removal efficiencies across the studies.Experimental variables include CWS type, pollutant concentration and form, adsorbent dosage, hydraulic retention (contact) time, pH, and temperature.
The data are also shown as box plots, highlighting data ranges and trends across the studies (Figure 3).The initial pollutant concentration(s) (Figure 3a), CWS adsorbent dosage (Figure 3b), HRT (Figure 3c), and wastewater volume (Figure 3d) for all CWS types are compared.Abbreviations for each shell type are denoted as oyster shells (OSs), mussel shells (MSs), eggshells (ESs), clam shells (CSs), white hard clams (wCSs), bivalve shells (BSs), zebra mussels (ZSs), and combination of shells (CoS).Figure 3a displays the initial pollution concentration of phosphate (P), ammonium (A), and nitrate (N) for different shell types.Generally, a larger nutrient concentration range (2-1500 mg/L) was observed for the main shell types studied (i.e., OS, MS, and ES).Phosphate removal in OS was tested between 2 and 600 mg/L with a median of 80 mg/L.In contrast, MS showed two distinct ranges, 5-20 mg/L (low) and 150-1000 mg/L (medium-high), with an overall 10 mg/L median.ES had a higher median of 400 mg/L, with more studies conducted at higher phosphate concentrations.Higher phosphate concentrations in synthetic wastewater compared to non-synthetic wastewater were used for all shell types to support isotherm and kinetic studies [8,36,91,92].Conversely, when lower phosphate concentrations were used (in MS), they corresponded to phosphate concentrations typical in freshwaters (e.g., river or lake) (Table 1).
The median ammonium concentration tested in OS was 60 mg/L, while MS (median 300 mg/L) showed three distinct concentration ranges between 2 and 500 mg/L.The high ammonium concentration corresponded to synthetic wastewater, raw landfill leachate, and swine wastewater.The medium ammonium range (100 mg/L) occurred in synthetic and domestic wastewater, while the low ammonium level (<10 mg/L) was from lake water and aquaponic wastewater (Table 1).Very few studies were observed using OS and ES for nitrate removal, with none recorded for MS.The nitrate concentration tested for OS ranged from 6 to 300 mg/L with a median of 200 mg/L, while for ES, it ranged higher at 80-800 mg/L (median 400 mg/L).Like ammonium, a high initial nitrate concentration prevailed in synthetic wastewater, with a medium-strength nitrate concentration in domestic wastewater and a low nitrate concentration in aquaponic wastewater and treated effluent.
Figure 3b shows the range of adsorbent dosages (in g/L) for different shell types.A larger range of adsorbent dosages was found in studies using OS, MS, and ES (0.1-100 g/L), a reflection of the greater number of studies using these shell types.While their medians are relatively close (10-20 g/L), the mean for OS (90 g/L) was higher than those of MS and ES (both at 25 g/L), suggesting that OS has a wider application in water treatment (represented by an extended distribution plot).Oyster shells dominated as a CWS adsorbent compared with other shell types.This may be because of their wider availability, higher calcium content, and higher resulting alkalinity.As stated earlier, the CaCO3 content in oysters, mussels, and eggshells ranges from 90 to 97%.Interestingly, the mineral composition of shells from the same species but collected from different places also differed [15], which might be attributed to variations in shells' nutrition and age and the habitat's temperature, salinity, and pH [96].Furthermore, differences in pyrolysis conditions likely affect the dosages used in the experimental studies.For example, CWSs produced by slow calcination have a higher BET surface area (81 m 2 /g) than when produced by fast calcination (20 m 2 /g) [46].
Other shell types, such as Zebra mussels, clams, white hard clams, bivalves, and combined shells, were studied in a narrower adsorbent dosage range, a reflection of the fewer publications and possibly shell regional availability.For example, abundant bivalve shells in the freshwater lake of Tonle Sap, Cambodia, were used as adsorbents for phosphate removal to control eutrophication [55], and white hard clam waste shells were utilised to remove phosphorus from anaerobically digested swine wastewater in Vietnam [56].Zebra mussels, an invasive species in North America, were mined for calcium carbonate to remove phosphorus from domestic wastewater [31].
The influence of hydraulic retention time (HRT), or the time the CWS adsorbent was in contact with the wastewater, on each shell type is shown in Figure 3c.Most shell types were investigated within an HRT range of 0.083-50 h.Median values for all shell types ranged from 10 to 20 h, except for combined shells, which had a higher median of 30 h.The two highest data points in the CoS studies corresponded to an unusually long contact time of 4 and 15 days in batch adsorption assays that skewed this median [5,94].
The wastewater volume used in the experiments is shown in Figure 3d.OS and MS sample volumes ranged between 20 and 600 mL, with median values ranging between 50 and 100 mL (like ES), suggesting that most laboratory experiments considered 50-100 mL sufficient to investigate the effect of CWM on nutrient removal under laboratory conditions.The exception was a study that reported combined shells with a median of 300 mL used in the experiments.Field trials using CWS-derived media to treat nutrients would require much greater volumes of wastewater (synthetic or real) and provide valuable additional datasets to understand the effects of wastewater volume on nutrient removal.

Isotherms and Kinetics
For the practical application of adsorption processes to treat pollutants, it is crucial to have data related to equilibrium, the adsorbent's adsorption capacity, and the kinetics of the process [97].Adsorption kinetics provide important data for designing, optimising, and troubleshooting treatment systems relying on adsorption.Adsorption isotherms explain the adsorption/desorption mechanisms of particular treatment processes [66].Table S2 (Supplementary Materials) summarises the removal mechanisms, isotherms, and reaction kinetics reported in the experimental studies reviewed.Figure 4 shows the CWS type, and the isotherm and kinetic models reported.
Calcination causes volatilisation of the organic components, increasing the calcareous shell's specific surface area and porosity [19,33].Heating the shell to above 500 °C converts CaCO3 to CO2 and calcium [28] and changes the crystal structure of CaCO3 [6].In this stage, the orthorhombic crystal structure-or aragonite, the main form of calcium carbonate in CWSs-is transformed into a calcite polymorph with a trigonal rhombohedral structure [4,6,101].Calcination is the most widely used method to enhance CWS capacity to remove nutrients from wastewater (Figure 4a).Most papers reporting calcination or other CWS modification used oyster shells, reflecting the dominance of OS in the experimental studies.
CWS acidification is applied by immersing the shells in a low-concentration acid solution overnight to dissolve CaCO3, thus producing CaO [89].The hydrothermal method involves crystallising minerals from high-temperature aqueous solutions in an autoclave or a microwave under high vapour pressures [102].Lai [78] processed oyster shells by sintering with silica micro-powder using high-temperature and observed that hydrothermal treatment showed a large adsorptive phosphorus capacity and long 'service life'.The sintering and hydrothermal treatment result in the partial transformation of CaCO3 into CaO, which partly reacts with SiO2 (silicon dioxide) to produce CaSiO3 (calcium silicate), resulting in activated Ca 2+ which reacts with phosphate in wastewater to form calcium phosphate precipitates [76,78].Isotherm models reported in the CWS studies are shown in Figure 4b.Common models (in descending order of frequency) are Langmuir, Freundlich, and Langmuir-Freundlich.The Langmuir and Freundlich models are the two most common adsorption isotherm models to describe adsorbate removal processes and mechanisms [76].While the Langmuir isotherm model involves adsorbate interaction on completely homogenous surfaces with the same activation energy, the Freundlich isotherm model is suitable for highly heterogenous surfaces [34,76].The Langmuir-Freundlich, or Sips, isotherm assumes the surface is homogeneous, but the adsorption is a cooperative process due to sorbate-sorbate interactions.This model acts like the Langmuir model (monolayer adsorption) at a high adsorbate concentration, while it behaves like the Freundlich model at a low adsorbate concentration [92,103].Other relevant isotherm models reported in the CWS adsorption studies are Temkin [69], Dubinin-Radushkevich (D-R) [95], and Koble-Corrigan [58].
It was found that two-thirds of the reviewed studies involved isotherm investigation.Langmuir is the most common isotherm in studies involving oyster shells, mussel shells, and eggshells, indicating that adsorption occurs by monolayer deposition at the surface, with each site only allowing one adsorbate unit.The Freundlich isotherm was more commonly reported in studies with mussel shells, which suggests multilayer adsorption on non-uniform surfaces on this type of shell [103].It is postulated that active sites on the mussel shells' surface strongly bind with the first layer of the adsorbate (e.g., phosphate), and that the binding strength decreases with increasing adsorbate layers.Furthermore, oyster shells are mostly used in raw form (Figure 4); for this, the Langmuir isotherm seems to produce the best fit.In the case of mussel shells, they are mainly calcinated, and the Freundlich model seems to work best.Interestingly, a Langmuir-Freundlich isotherm is only observed for eggshell studies, implying that monolayer and multilayer adsorption occur simultaneously.
Of all 64 reviewed studies, only half reported process kinetics data (Figure 4c).Of these, most (73%) reported the pseudo-second-order (PSO) model to explain the experimental data better than the pseudo-first-order (PFO) (18%), Elovich (6%), or Avrami (3%) models.Nutrient removal by the three main shell types (i.e., oyster, mussel, and eggshells) followed PSO rather than PFO kinetics.The PFO model generally fits the data well over the first 20-30 min of the adsorption process, while the PSO model describes well the adsorption behaviour over the whole time range [103].Hence, it is unsurprising to note a high agreement between the PSO model and the Langmuir isotherm [7,65,103].The PSO kinetic model assumes that the adsorbate removal process is controlled by chemisorption with predominant monolayer adsorption [7,20].In contrast, the PFO model assumes that the adsorption rate is determined by physical interaction with a one-on-one adsorption mechanism in which one pollutant unit is adsorbed onto one adsorbent surface site [8].The Elovich kinetic model is only observed for clam shells, suggesting adsorption on very heterogeneous shell surfaces [58] and possible chemical reactions on the surface of the shells (adsorbent) [57].

Removal Mechanisms
Raw CWSs are mainly made of CaCO3 aragonite [6,60], which is less effective as an adsorbent for phosphate removal.Upon calcination at 500-600 °C, CaCO3 aragonite converts to calcium carbonate calcite (also known as calcite) [4,6,60], and further calcination at 800 °C produces CaO, the most effective form for phosphate precipitation [25,33,38].It is reported that when using CWSs, phosphate is removed by retention on calcite rather than on aragonite [4] via adsorption and the nucleation of calcium phosphate (hydroxyapatite) precipitate [6].Adsorption removal on calcite occurs on the particles' surface [4], creating sites for the initial binding between calcium and phosphate [69] and subsequently initiating the heterogeneous nucleation of calcium phosphate precipitates on calcite surfaces [6].With time, more precipitates are formed by the reaction of hydrated CaO or CaOH with phosphate [60], thus enhancing phosphate removal through a precipitation mechanism [4].It was found that precipitation can remove phosphate in larger stoichiometric amounts than adsorption [4].
In line with the above, Abeynaike et al. [6] reported that a 95% or higher phosphate removal was achieved with partially calcinated mussel shells (calcite) compared to 22-45% with raw mussel shells.A similar result was also obtained by Daudzai et al. [60], where 85% phosphate removal and no nitrate removal were recorded for unrinsed raw seashells.In contrast, unrinsed calcinated seashells were able to remove almost 100% of phosphate and nitrate.
Nitrogenous compounds, such as ammonium (NH4 + ) and nitrate (NO3 − ), are removed via chemisorption [18,35,36] or physisorption [10].Monovalent nitrate anions (NO3 − ) are known to bond weakly with divalent calcium cations (Ca 2+ ) compared to trivalent phosphate anions (PO4 3− ) [60].In addition, their high solubility makes them difficult to remove.However, the presence of O-Ca-O functional groups and CO3 2− ions in calcinated mussel shells facilitate nitrate removal by the formation of calcium nitrate-hydroxide complexes [60], as shown in Figure 5. Ammonium removal via chemisorption reactions such as ion exchange and electrostatic attraction occurs heterogeneously on multilayers of raw mussel shells, with two dominant functional groups, C-O and C=O, playing the main role [18].
While the higher presence of the C-O functional group and inorganic CO3 2− in calcinated eggshells improves ammonium adsorption, the specific surface area did not significantly influence ammonium adsorption [35].

Effects of Experimental Variables on Nutrient Removal
The effect of CWS adsorbent particle size (Figure 6), adsorbent dosage (Figure 7a), and hydraulic retention time (Figure 7b) on nutrient removal was further investigated in the form of scatter plots and correlation analyses.Nutrient removal showed no clear correlation with CMS adsorbent particle size (Figure 6a (phosphate) and Figure 6b (nitrate and ammonium)), probably due to the large variation in experimental conditions used across the different studies.These variables, such as different wastewater types, pollutant strength, raw or modified CWSs, and operational conditions (pH, temperature), likely affect the adsorption capacity regardless of the particle size.More data were available for phosphate than nitrogen compounds to draw any potential relationships.Oyster shells had a larger particle size range with strong adsorption capacity (up to 1000 mg/g) across this range (Figure 6a), suggesting that they are an effective adsorbent for phosphate removal from wastewater [39].In contrast, eggshells exhibited similar adsorption capabilities as oyster shells but were only tested with particle sizes of <1 mm.Mussel shells were tested within a much smaller particle size range (around 1 mm) but showed very large variations in adsorption capacity.The relationship between pollutant removal efficiency and adsorbent dosage across the studies is shown in Figure 7a,b (insufficient data were reported for nitrate to include).The only CWS type that showed a relationship between adsorbent dosage and treatment efficiency was oyster shells (n = 4) treating ammonium at a relatively high adsorbent dosage (Pearson's r: 0.9249, R 2 = 0.8555).Nonetheless, eggshells consistently removed >80% phosphate, with both oyster and mussel shells' phosphate removal being more variable (between 20 and 100%) (Figure 7b).One important variable to note is that most of the oyster shell studies used synthetic wastewater compared to the mussel shell studies, which may help explain the greater MS treatment variability.
The relationship between nutrient removal and HRT is shown in Figure 7c,d.There were insufficient (or absent) data reported in the reviewed studies to comment on the performance of some CSW-nutrient combinations (e.g., ammonium removal by eggshells and nitrate removal with oyster and mussel shells).The only CWS types that showed apparent relationships between the parameters were oyster shells treating nitrate (Pearson's r: −0.9374, R 2 = 0.8787) and mussel shells treating phosphate (Pearson's r: 0.8047, R 2 = 0.6476), revealing that a higher nitrate removal is achieved with a lower HRT for oyster shells, and a consistently high phosphate removal at >90% for mussel shells regardless of the HRT (indicative of a fast and relatively stable reaction) (Figure 7d).

Possible Reuse of Post-Treatment Adsorbents
A few studies reviewed in this manuscript discussed nutrient desorption, particularly phosphate, after wastewater treatment [60,69,104].Desorption would allow for nutrient recovery and reuse and would help expand the lifespan of the adsorbents as they are reused [69].The desorption of phosphate was performed using 2% citric acid for an hour with more than an 80% successful regeneration rate [69].Daudzai et al. [60] reported the use of 0.5 M HCl at a 1% dosage, achieving 80% and 98% successful regeneration for phosphate and nitrate, respectively.Coincidently, both chemicals (2% citric acid and 0.5 M HCl) were applied as desorption agents with a high regeneration rate above 90% [104].The full desorption of phosphate occurred with 2% citric acid and 0.5 M HCl, suggesting that the adsorbed phosphate would be bioavailable to plants if the CWS adsorbent was to be used as a soil conditioner for agriculture [104].Pap et al. [104] also suggested that posttreated CWS adsorbent could be potentially applied as a soil amendment in acidic soils with limited phosphate-adsorbing capacity.Another possible reuse of spent CWSs would be as construction material, e.g., a replacement for magnesium phosphate cement [48].

Conclusions
This systematic review on nutrient removal from wastewater by calcareous waste shells (CWSs) revealed that most studies were conducted in the laboratory using synthetic wastewater and small (50-100 mL) wastewater volumes.Studies have focussed more on phosphorus removal than on nitrogen.Much higher phosphate removal levels have been reported compared to any nitrogenous compound.It also revealed that calcination is the most frequently used method to enhance nutrient adsorption capacity.In terms of sorption models, the Langmuir isotherm fits experimental data with raw oyster shells best, while the Freundlich model fits data better from calcinated mussel shells' experiments.Furthermore, a higher agreement has been reported between experimental data and the pseudo-second-order kinetic model (73%) than the pseudo-first-order model (18%) for all shell types.An examination of the effect of individual experimental variables on nutrient removal was inconclusive, mainly due to the limited number of individual data points reported coupled with the confounding effects of multiple different experimental variables across the studies.Identifying and controlling for confounding variables are crucial in establishing reliable correlations and drawing meaningful conclusions from research findings.The application of laboratory results to real-world solutions requires certainty on the effects of critical individual variables such as HRT, temperature, and wastewater strength.While there is commendable laboratory research showcasing the valorisation of CMS in wastewater treatment, this needs to be translated into larger-scale trials to provide more certainty in sustainable design solutions and a pathway for scaling with certainty in effectiveness.
Several potential research questions arise that can be explored in future studies: 1. Does the CaCO3 content of different CWSs affect nutrient removal? 2. Which experimental variable (HRT, adsorption dosage, particle size, or wastewater strength) is most influential on nutrient removal capacity? 3. To what extent would climatic variables influence the adsorption capacity of CWS removing nutrients if experiments were performed under field conditions? 4. Would the isotherms and kinetic studies change if studies were conducted for longer under field conditions? 5. Can CWSs be modified (functionalised) differently to improve nitrate adsorption?Supplementary Materials: The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/environments11060119/s1.Table S1: Experimental conditions and removal efficiencies using CWS; Table S2: Isotherm and kinetic results of studies that used CWS for treating nutrients in water.

Figure 1 .
Figure 1.Literature retrieval and screening process for the review (97% journal articles and 3% conference papers).

Figure 2 .
Figure 2. Number of publications by (a) country, (b) year, (c) wastewater type, and (d) nutrient forms.Total (n) of publications = 64 (Figure 2a,b); 59 for P and 23 for N (Figure2c); 81 (Figure2d).The total number of studies (81) across all wastewater forms in Figure2dis greater than the total number of publications(64) because some studies reported multiple nutrient pollutant forms within the same study.Palestine (1) refers to 1 publication.

Table 1 .
Studies using calcareous waste shells (CWSs) to treat nutrients in wastewater.