**3. Properties and performance of GPs/AAMs towards various contaminants**

Despite the fact that the first identification of GPs as unconventional construction materials was in 1979 [158], broader applications of GPs/AAMs started in late 90s. Although GPs/AAMs are to be considered by some authors as an economic alternative to zeolites or activated carbons for water purification, the lack of real cases reported is obvious. To urge commercial importance, GP/AAM adsorbents should be readily available, economically feasible, steady in characteristics, and easily regenerated. Several comprehensive reviews on the GP/AAM materials for the water treatment sector have been published just recently [57, 150, 151, 153]. Therefore, in this section the bright and promising works will be highlighted as well as challenges and trends for future studies revealed.

*GPs/AAMs for metal(oid)s removal.* The adsorption characteristics of individual species and particular conditions of adsorption could be found elsewhere [57, 150, 151, 153]. Here, we would like to emphasize some challenges and gaps, which might be addressed in future studies. There are only several articles discussing selectivity of adsorption on GPs/AAMs alongside the matrix effects. In most of the studies pure mono-element aqueous solutions were implied, and the adsorption characteristics for individual substances- without possible influence of matrix macro-elements have been established. However, GPs/AAMs that are considered as a replacement of zeolites have to demonstrate selectivity under complex matrices in order to have opportunities to promote the implementations in various industrial applications.

In order to obtain adequate adsorption parameters, an excessive alkaline residue in GP/AAM should be washed out properly (pH 7 0.5 within 24 h required) [159]. Otherwise, the increment of pH of aqueous solutions containing heavy metals will favor the hydroxide precipitation process, leading to wrong result interpretation. For porous GPs, washing away the excessive alkalis resulted in the increment of total porosity [11], which led to better performance. Moreover, excessive alkalis were used intentionally to neutralize AMD [42] and remove metal ions. However, a strict protocol must be followed to characterize newly designed materials.

Selective adsorption is relies on several factors such as a metal ion activity, hydration radius and free energy of hydration, and a pore size distribution of GP.

Geopolymerisation by itself could lead to the formation of new ion-exchange sites at the GP surface, but additives in composite formulations could have even higher influence the adsorption characteristics.

An ionic exchange reaction between the heavy metal ions and sodium ions has resulted in heavy metal removal by the metakaolin GP [159]. The adsorption selectivity of heavy metal ions by the GPs at pH 4 in multi-component solution was in the following order: Pb2+ > Cd2+ > Cu2+ > Cr3+, while qe [mg/g]: 100 > 76 > 55 > 10. The order of adsorption was in accordance with the hydrated radius and free energy of hydration for selected ions. However, the free energy of hydration and the activity for Cr3+ are all higher compared to those of other metals, though its adsorption rate does not correspond to the assumed order. The selectivity towards Cr3+ was be explained through its ionic status. When the pH exceeded 4, Cr3+ transforms to Cr (OH)2+, which might lead to its lower adsorption ability. It is also noted that at lower pH, the balancing ions present on the GP surface tend to be replaced by the hydrogen ions instead of the metal ions that lead to lower capacity at acidic pH.

Lopez et al. [5] investigated the selectivity of metakaolin-based GPs in multicomponent solutions (Pb2+, Cu2+, Cd2+, Ni2+, Zn2+ and Cs<sup>+</sup> ). For a composition with Si/Al ratio 2, the best capacities and selectivity towards Pb2+ and Cs+ were observed. The adsorption selectivity for the mixture of metal ions was in the following order Cs<sup>+</sup> > Pb2+ > Cu2+ > Zn2+ > Ni2+ > Cd2+, while qm [mg/g]: 43 > 35 > 15 > 3 > 1 > 2. The adsorption capacity for individual elements were higher: 57 mg Pb2+/g > 52 mg Cs<sup>+</sup> /g > 46 mg Cu2+/g > 14 mg Cd2+/g > 9 mg Zn2+/ g > 4 mg Ni2+/g. Moreover, the effect of solution salinity (NaCl, 5% and 10%, wt) was studied, and no considerable effect on the adsorption order of metal ions or GP capacity in multi-composition solution was found. The authors presumed the existence of at least two types of binding sites with different affinities toward the metal ions to explain such a tolerance.

Selectivity of GP composites with zeolite filler was studied by Andrejkovičová et al. [4]. The highest adsorption was observed for Pb2+ for all the GPs obtained, while an adsorption order was as follows: Pb2+ > Cd2+ > Zn2+ > Cu2+ > Cr3+. The adsorption of Cu2+ and Cr3+ increased as the amount of metakaolin in the GP increased, whereas the composite with 25% zeolite doping had higher adsorption characteristics towards Pb2+, Cd2+ and Zn2+. GPs prepared from zeolitic tuff and kaolinitic soil by El-Eswed et al. [160] showed totally different order of adsorption: Cu2+ > Pb2+ > Ni2+ > Cd2+ > Zn2+. Moreover, the adsorption order strongly depended on the GP composition, although Cu2+ and Pb2+ adsorption has always prevailed.

The ability of BFS- and metakaolin-based GPs to remove Ni2+ and metalloids (As and Sb) in form of oxyanions was shown in [32]. Both adsorbents completely removed Ni2+ that most likely was associated with precipitation of its hydroxides on the GPs, while both metalloid oxyanions were adsorbed by BFS-GP equally. Another remarkable merit is that the adsorption capacities were obtained with real matrixes (spiked mine effluents), and were 4.42 mg/g, 0.52 mg/g, and 0.34 mg/g for Ni2+, As3+, and Sb3+, respectively. It is specified by the authors that the low capacities could be a result of competition of some matrix ions (Sr, Ca, Mg, Mn) with the target ions for binding sites.

Researches with increasing frequency pay attention to this problem and try to demonstrate the removal efficiencies with real samples. Removal of Ca2+ and Mg2+ from intact groundwater was examined in [58] on kaolin-based GP. With adsorbent dose of 1 g/L, the removal rate were 37.5% and 16.2% for Ca2+ and Mg2+, respectively. Metakaolin-based GP was tested by Kara et al. [87] for Mn2+ and Co2+ removal from real wastewater. The removal rates in real wastewater decreased from 97.5% to 53.01% and 94.6% to 39.12% for Co2+ and Mn2+, respectively. The results demonstrated that the adsorption performance affected negatively by the coexistence of some other cations and/or anions in the adsorption medium. Bentonitebased GPs were used for heavy metals removal from synthetic wastewater [61]. Porous biomass FA-based GPs were used in [129] for simultaneous removal of heavy metals from wastewater samples. Mixed FA/metakaoline-based GPs were used in [103] for Cu2+ removal from real wastewater. In the showcase, the adsorption capacity of GPs towards Cu2+ decreased by 27% as compared to synthetic samples. Sithole et al. treated acidic industrial effluents by FA/BOFS-based GPs [42, 43]. New GPs containing hollow gangue microsphere were applied for Zn2+ removal from smelting plant wastewater in [93]. At an adsorbent dose of 30 g/L, a complete Zn removal was observed. The distinctive aspect of the reported cases was that a complex composition of treated solutions is likely to decrease substantially capacity of the GP. Thus, the adsorption capacities obtained for the ideal laboratory conditions should be primary used as the guiding not decision-making parameters.

*GPs/AAMs for removal of other inorganic ions.* Besides metal(oid)s, GPs/ AAMs were examined for removal of ammonium and various anions. Luukkonen et al. [149, 152, 161] showed potential of metakaolin-based GPs to remove ammonium. The optimized GP composition was proposed and manufactured in both powder and granular forms. The efficiency of removal was demonstrated in municipal wastewaters (primary and secondary effluents) as well as landfill leachates. Metakaolin-based GPs prepared from commercial and waste metakaolin were able to effectively remove ammonium from synthetic and wastewater samples [59]. In fact, GPs prepared from paper mill fiber sludge showed better selectivity in the presence of competing ions under real matrix conditions. Bai and Colombo prepared metakaolin-based GP foams in the form of monolithic porous filters [162, 163]. The filter was able to remove up to 95.3% of ammonium from runoff waters at the initial concentration of 3 mg/L.

The removal of phosphorus was attempted in [10] with a pervious FA-based GP. The removal rate increased with the increase of pH. Up to 85% of phosphorus were removed from a treated wastewater. Simultaneous removal of ammonium and phosphate by composite metakaolin/BFS-based GPs was demonstrated in [91]. Phosphate removal was enhanced in presence of ammonium. At slightly alkaline conditions (pH 7– 8), the removal rate towards phosphate ions was relatively high (>86%), whereas the ammonium removal up to 35% was also achieved. FA-, BFS- and fiber sludge GPs were investigated as promising adsorbents for phosphorous removal from diluted solutions. The capacities at initial phosphate concentration of 100 mg/L are 26 mg PO4/g for BFS-GP, 36 mg PO4/g for FAF-GP, and 43 mg PO4/g for FSHCa-GP [115].

Sulfate ions were removed by barium-modified BFS-based GPs [134]. Adsorption capacities were 91.1 and 119.0 mg SO4/g for model solution and mine effluent, respectively. The surface complexation or precipitation of barium sulfate were suggested as probable removal mechanisms.

Removal of halides by GPs/AAMs is an emerging topic. For this end, composite or functionalized materials are designed. Removal of F ̶ was demonstrated by slag-based GP microspheres modified with CeO [138], Fe2O3 [136], and bivalent metallic species [41] with capacities towards the contaminant 127.7 mg/g, 59.8 mg/g, and 60 mg/g (zinc impregnated BOFS-GP), respectively. A metakaolin-based GP functionalized

by surfactant was developed for efficient removal of radioactive iodide [97]. High concentrations of competitive anions had limited influence on the adsorption process.

*GPs/AAMs for removal of organic substances*. In fact, GPs contained residual metal oxides could have potential catalytic performance. Thus, GPs based on industrial by-products such as FA, BFS, or their mixtures with silica fume and aloxite demonstrated catalytic activity under visible light irradiation. The descriptive list of organic substances removed by GPs could be found in reviews [57, 150, 151, 153]. Manly, cationic and neutral dyes were investigated as targets in recent studies, although removal of fecal coliforms [10], volatile organic compound [77], and tetracycline [164] was reported.

Oxidative degradation or photodegradation after adsorption have been specified by authors as primary mechanisms of organic pollutants' removal. Although conventional GPs have been reported for these purposes [86, 89, 104, 126, 139], they would rather have had low adsorption/degradation characteristics. Hybrid or composite materials were proposed to improve the removal efficiency of organic pollutants. Thus, graphene [120, 132, 133, 165], TiO2 [88, 98, 105], CdS [142], various metal oxides [101, 106, 135] were introduced in GP matrix in order to enhance degradation abilities of resulting materials.

### **4. Regeneration of GPs/AAMs and further resource recovery options**

In last a few decades, significant improvements were made in both efficiency and economy in removal of metal(oid)s and other substances by adsorbents. Nevertheless, regeneration and recycling of used adsorbents, or recovery of the removed species from the desorbing agents are still rarely reported. For regeneration and reuse of GPs/AAMs, various possible regenerating agents such as acids, alkalis and chelating agents could be used. Only a few of the reported studies were focused on recovery of adsorbed (from saturated adsorbents) and desorbed (from regenerating agents) metals [11, 87, 96, 131]. However, for industrial application and success completion of new GP/AAM adsorbents on the market, research studies on number of adsorption–desorption cycles are in high demand. Moreover, revenues gathered from resource recovery options will have a decisive role in further technology implementation.

The regeneration of metakaolin-based GP by sodium chloride under alkaline conditions after ammonium adsorption for the first time were demonstrated in [152]. Three adsorption–desorption cycles were carried out with a steady removal efficiency. Sodium chloride and sulfate, potassium sulfate and phosphate were studied in [59] as regenerating agents for saturated metakaolin-based GPs. Sodium sulfate showed better results during five cycles under continuous sorption–desorption experiment, only 34% of an initial overall capacity of the GP were lost. Sodium chloride regenerant was also efficient, but only 55% of ammonium could be removed after 5th desorption cycle. The same adsorbents were used to test a nitrogen recovery option in a laboratory-scale demonstration setup [166]. The layout consisted of an adsorption/desorption unit and Liqui-Cel® membrane. A liquid phase obtained during adsorbent regeneration was purified in the membrane contactor in order to recover ammonium nitrogen as ammonium sulfate or phosphate. The purified regeneration solution was used repeatedly for further adsorbent regeneration. Several regeneration-purification cycles were conducted to estimate system sustainability and chemical consumption demand. Operational conditions of a membrane process such as shellside and lumenside feed flows, temperature, and pH were adjusted to gain maximal capacity of the setup. One membrane contactor (2.5 8-inch Liqui-Cel) was used under following operational conditions: 100 L/h shellside and 60 L/h lumenside feed flows, 40°C working temperature, pH ≥ 10.

Technical sulfuric or phosphoric acids, up to 5%, were used as lumenside phases. The concentration of ammonium-content salt in a resulting received phase were 17% and 22% for phosphate and sulfate salt, respectively.

Metal recovery from GPs/AMMs via ion-exchange mechanism can only take place if physical adsorption occurred and the pH was low enough to prevent precipitation of metal hydroxide during adsorption process. Acids of over 0.1 M strength affect the structure of the GPs, and while metals are regenerated by acid washing, the reuse of adsorbents are diminished both in batch [11] as also in continuous mode [87, 167] experiments. Mild acid washing with 0.01 M H2SO4 or HNO3 removed metals from GPs efficiently in short time (1–2 h). It has also been shown that the adsorption capacity after mild acid washing could increase [131], which could be explained by exchange of Na+ with easier replaceable H+ cations. Selective desorption of copper has been observed by ammonia. A linear desorption ability with respect to ammonia concentration was observed, and complete desorption being possible by 10% ammonia solution [50, 61].

Sequential desorption tests of Cd2+ have been conducted on a loaded metakaolin GP, establishing the percentages of physically adsorbed, ion-exchangeable, EDTA extractable, and residual forms of metal [96]. The authors showed that physical adsorption is negligible, and ion-exchange with MgCl2 constituted to only 2–8% of adsorbed Cd2+. The bulk amount of Cd2+ adsorbed by the metakaolin GP was EDTA extractable, and the adsorbent remained 85% of its adsorption capacity after EDTA desorption for 5 cycles. Luukkonen [32] and Naghsh [58] suggested the efficient metal desorption by 5% NaCl. However, care must be taken since the balancing ions can form a positively charged film on the adsorbent surfaces. El Esweed et al. have achieved ion-exchange based desorption of Cu2+ by 0.1 M NaCl [160]. From all the studies reported, only Cd2+ has been shown to be desorbed at pH > 8 with NaOH solution, achieving 24–84% desorption [64].
