Cross-linked-substituted (esterified/etherified) starch derivatives as aqueous heavy metal ion adsorbent: a review

Heavy metals are naturally occurring chemical elements with high atomic numbers, and densities at least five times higher than that of pure water. Another description states their atomic mass and atomic number to be higher than 23 and 20 respectively, and density greater than or equal to 5 g/cm³ (Tchounwou et al. 2012; Koller & Saleh 2018). Some heavy metals (micronutrients) such as copper, iron, cobalt, selenium, manganese and nickel (in trace amounts) are essential to human health but toxic above some threshold concentrations, while others, e.g. cadmium, mercury, arsenic, lead, thallium and chromium, are non-essential and toxic at any level (Rengel 1999; Sharma & Agrawal 2005; Tchounwou et al. 2012; Ali & Khan 2018; Koller & Saleh 2018). While some heavy metals are of outstanding technological significance (in the production of catalyst, batteries, alloys, pigments etc.), others (e.g. gold, silver etc.) are regarded to be precious and valuable in jewellery and carvings (Tchounwou et al. 2012; Oves et al. 2016; Koller & Saleh 2018). Therefore, uncontrolled discharge of heavy metals into the environment can be regarded as ‘waste of resources’ as much as it can be harmful to health. Careless discharges of wastewater from industrial and agricultural activities are the major sources of heavy metal pollution of environmental compartments such as soil and waters (Tchounwou et al. 2012; Sulaymon et al. 2014; Chen et al. 2019; Shalla et al. 2019). This is typical for small ventures lacking the funds to install the facilities required for adequate treatment of these wastewaters before discharge. This situation keeps pressing for the design and development of cheaper techniques for the removal of heavy metals from their typical aqueous media.

Among the wastewater remediation techniques that have been successfully used are precipitation, electrolysis, ion-exchange, membrane filtration, electro-dialysis, reverse osmosis, complexation, oxidation–reduction process, evaporation, photocatalyic degradation/mineralization etc. (Ahalya et al. 2003; Sancey et al. 2011; Akinterinwa et al. 2016; Madala et al. 2017; Omotunde et al. 2018). However, in terms of cost and maintenance, adsorption techniques are gaining a huge preference. Adsorption is a surface phenomenon which may include entrapment of aqueous heavy metal ions onto the adsorbent via physical or chemical interactions or both. The cheapest adsorbents have been developed from numerous varieties of treated/modified biomasses which are industrial or agricultural by-products (Ahalya et al. 2003; Abdel-Raouf & Abdul-Raheim 2017). Starch derivative adsorbents fall under this category; however, they are outstanding because their modification can be tailored to achieve specific adsorption functionalities.

Starch is natural, cheap, readily available, renewable, and biodegradable (Cummings et al. 2019). Native starches are biopolymers, and the physicochemical properties may vary based on the botanical sources, thereby providing natural alternatives for applications (Klimaviciute et al. 2010; Ashogbon & Akintayo 2014; Reddy et al. 2017). Hydroxyl and acetal oxygen functional groups in the native starch (Figure 1) constitute potential adsorption sites (Rendleman 1978; Vandenbossche et al. 2015; Salah et al. 2019). Hence, native starches have been used as aqueous heavy metal ion (AHMI) adsorbents among other applications (Hood & O'Shea 1977; Ekebafe et al. 2012; Soto et al. 2016).

Typically, industrial application of the native starch is limited due to some physicochemical and structural failure under application conditions. Its outstanding economic and environmental attributes, however, drove scientific and technological innovations to remedy these limitations via tailored modification processes yielding starch derivatives with novel properties. Starch modification can be defined as an induced and tailored change in the properties of the native starch to suit an intended application better. There are numerous possibilities of starch modifications, and these have presented a wide spectrum of starch applications in food and non-food productions (Ashogbon & Akintayo 2014; Alcázar-Alay & Meireles 2015; Haroon et al. 2016). Starch modification can be defined as an induced and tailored change in the properties of the native starch to suit an intended application better. Basically, modification of starch can be categorized as; genetic, enzymatic, physical and chemical modification processes. Genetic modification relies on bio-engineering of plants to synthesize starch with specific properties. High amylose content (about 70%), waxy starches (with 99–100% amylopectin content), modified amylopectin structure starches, and starches with modified granule size and amount have been produced by genetic modification (Kaur et al. 2012). Enzymatic modification was achieved by subjecting native starch to enzyme activities. Enzyme (amylases) hydrolysis of starch has yielded glucose syrups and maltodextrins (van der Maarel et al. 2002; Nitschke 2009). To achieve some physical modifications, native starch was subjected to conditions within which temperature, pressure, shear, moisture, irradiations or a combination of these is controlled to obtain products with new properties (Ashogbon & Akintayo 2014; BeMiller & Huber 2015; Schmiele et al. 2019). Chemical modification entails tuning the chemical composition and functionality of the native starch to obtain new products with desired properties via reactions with some chemical reagents. Typically, OH groups on starch molecules are attacked, resulting in replacements (substituion reactions) and/or inter-/intra-molecular bonding (cross-linking reactions) (Moorthy 1985; Kuakpetoon & Wang 2001; Zia-ud-Din & Fei 2017; Haq et al. 2019). Dual modification is achieved with different combinations of the categories earlier mentioned, for example physical–chemical (e.g., acetylation assisted by microwave, phosphorylation assisted by high pressures) or combinations of two different methods under the same category, for example chemical–chemical (e.g., acetylation/oxidation, cross-linking/acetylation, or cross-linking/hydroxypropylation) (Ashogbon & Akintayo 2014; Alcázar-Alay & Meireles 2015; Zia-ud-Din & Fei 2017).

Dual modifications have been identified to yield starch derivatives with structural, mechanical and chemical properties suitable for an efficient trapping of heavy metal ions in aqueous media via one or combinations of such mechanisms as; chelation, ion-exchange, electrostatic, molecular bonding and pore diffusion (ultrafiltration) (Rayford et al. 1979; Kim & Lim 1999; Wang et al. 2012; Xiang et al. 2016; Chen et al. 2019). Typically, one modification to enhance structural and mechanical stability in starch molecules is achieved via cross-linking, while the other modification substitutes OH with larger electron cloud functional groups such as xanthate, carbonyl, carboxylate, carbamate, acrylate, acetyl and phosphate (Figure 2), with stronger heavy metal ion adsorption affinities. Dual cross-linking–substitution modification procedures also allow better control on steric and conformation effects of the constituents of starch derivative adsorbents (SDAs) to enhance adsorption and adsorption selectivity (Ahalya et al. 2003; Deepatana & Valix 2008).

There is a rich literature on non-food applications of starch (Röper 2002; Ayoub & Rizvi 2009; Santana & Meireles 2014). Some studies have also been published on the use of starch derivatives as adsorbents (Carmona-Garcia et al. 2009; Haroon et al. 2016; Shalla et al. 2019). This review paper focuses on details of starch derivatives developed for adsorption of aqueous heavy metals via dual cross-linking-substitution modifications procedures categorized as; cross-link-esterification and cross-linked-etherification.

Cross-linking in starch can be initiated via physical (irradiation) and chemical means; however, chemical cross-linking is cheaper and simpler. Chemical cross-links stabilize the structure and mechanical activities of starch molecules and they can be likened to the muscles and tendons in the mammalian skeleton. Starch cross-linking occurs when bi- or multi-functional reagents (capable of forming ether or ester linkages with hydroxyl groups on starch), interact with starch macromolecules to form intra- and intermolecular bonding that creates three dimensional network structures (Figure 3) in the starch polymer matrix (Hirsch & Kokini 2002; Ayoub & Rizvi 2009; Alcázar-Alay & Meireles 2015).

Typically, small amounts of the cross-linking reagents are required to synthesize derivatives with optimum degree of cross-linking. This makes it a low cost procedure (Ayoub & Rizvi 2009; Shah et al. 2016). Examples of cross-linking reagents are epichlorohydrin (ECH), phosphorous oxychloride (POCl₃), sodium trimetaphosphate (STMP), sodium tripolyphosphate (STPP) and adipic–acetic mixed anhydride. (Hirsch & Kokini 2002; Koo et al. 2010; Ačkar et al. 2015). Primarily, the three dimensional network structures created in cross-linked starch derivatives prevent infinte granule swelling and dissolution over a wide range of conditions (e.g. temperature and pH) in aqueous media (Sancey et al. 2011; Braihi et al. 2014; Alcázar-Alay & Meireles 2015). The three dimensional network structure may also modify the pores on the native starch to create micropores capable of trapping AHMIs, thereby enhancing adsorption (Stancil et al. 2005; Xiang et al. 2016; Liu et al. 2018; Xie et al. 2019). For some cross-linking reagents, the functional group(s) on the cross-linkages possess electonic and/or valency attributes capable of interacting and trapping heavy metal ions in aqueous solutions. Li and co-workers carried out starch cross-linking with N,N′-methylenebisacrylamide (Figure 4) (Li et al. 2008). In addition to the structural and morphological modification, N,N′-methylenebisacrylamide also introduced chelating functional groups, i.e. carbonyl and acylamide, on the cross-linkages. The derivative obtained in the work was used in the adsoption of Cu²⁺, Co²⁺ and Ni²⁺, and the exothermic adsorption process improved with increase in metal ion concentration and reduction in aqueous solution temperature.

Another N,N′-methylenebisacrylamide SDA was synthesized by Yang et al. (2014). Results showed that adsorption of heavy metals studied was in the order Cu²⁺>Co²⁺>Pb²⁺>Cd²⁺. The adsorption behavior of Co²⁺ on the derivative was described by the first order kinetics equation, and was attributed to liquid phase mass transfer and pore diffusion (Yang et al. 2014). Bhat and co-workers synthesized POCl₃ cross-linked potato starch and used it in the adsorption of Cu²⁺, Ni²⁺, Zn²⁺ and Pb²⁺ (Bhat et al. 2015). After adsorption of ions, changes in surface morphology and formation of stable complexes (with metal–oxygen bonding) were reported in the study, and these were attributed to ion retention/trapping on the porous three dimensional polymer network and the interaction with P=O/P-O-C (on cross-linkages) respectively (Figure 5). The highest adsorption capacity of the cross-linked derivative (78% from 206 ppm aqueous solution) was obtained with Pb²⁺ (Figure 6).

Sago starch was treated with a mixture of mono- and di-sodium phosphate (3:2 mixture of 0.1 M Na₂HPO₄ and 0.1 M NaH₂PO₄) to synthesize the cross-linked sago starch phosphate, which was used in the adsorption of Pb²⁺ from aqueous solution (Irawadi et al. 2018). Cross-linking was confirmed with the emergence of the IR characteristic for phosphate diester bonding (RO-PO₃-R′), and the adsorption process was attributed to the complexation of Pb²⁺ with P=O and P–O (on cross-linkages), and –OH groups in the starch polymer matrix. The adsorption capacity was 23.67 mg/g, and the process was evaluated to follow second order kinetics and the Langmuir equation, hence suggesting a chemisorption process.

In these reports, starch derivatives used in adsorption were only cross-linked even though the cross-linking agent also introduced other functional groups with affinity for aqueous heavy metals. In the more common procedure, a second modification (substitution) is included to introduce specific functionalities meant to initiate or enhance the adsorption properties of the starch derivatives. There are many of these procedures; however, those that can be categorized as esterification and etherification will be discussed in the following topics.

Substituted starches are obtained from substitution reactions in which the hydrogen atoms on starch O-H groups are replaced by other molecules. The average amount of hydrogen atom substituted on the glucose units present in the starch macromolecule is analysed to determine the extent of substitution reactions, and the result obtained is referred to as the degree of substitution (DS) or molar substitution (Tessler & Billmers 1996; Fu et al. 2019; Abdul Hadi et al. 2020). Esterification and etherification are important examples of starch substitution reactions.

Esterification reaction occurs when some organic or inorganic reagents react with starch macromolecules and substitute the hydrogen atoms on hydroxyl groups. Organic starch esters are synthesized in reactions where the hydrogen atoms are substituted with bulkier molecules, creating ester functional group (St-COO-R) linkages with heavy metal ion chelation potential (Vandenbossche et al. 2015; Salah et al. 2019). Moreover, other functional groups on the bulk substituent may also enhance adsorption.

Esterifying organic reagents are mostly organic acids, their anhydrides and other derivatives (Figure 7) (Haq et al. 2019). For this reason, esterification reaction encompasses such common procedures as acetylation and succinylation (including octenylsuccinic anhydride modification) reactions (Lawal 2004; Ačkar et al. 2015; Won et al. 2017). Examples of some esterifying reagents are acetic acid, citric acid, oleic acid, maleic anhydride, vinyl acetate, adipic anhydride, octenylsuccinic anhydride, imidazolide of 3-carboxypropyl-trimethylammonium chloride, and betainyl chloride (Tessler & Billmers 1996; Zarski et al. 2016; Pfeifer et al. 2017; Goswami et al. 2020; Sun et al. 2020). Succinylated starch was synthesized and used in adsorption of AHMIs (Kweon et al. 2001). Adsorption capacity of the starch derivative increased with DS of the acetyl group, and the adsorption of heavy metals followed the order Pb²⁺ ˃ Cu²⁺ ˃ Zn²⁺ ˃ Cd²⁺ (Kweon et al. 2001).

Starch carbamate esters are synthesized when urea (carbamide) above its melting point (133 °C) reacts with starch following the esterification reaction pathway (Figure 8) (Khalil et al. 1994; Lewandowicz et al. 2000; Menzel et al. 2017). Starch carbamate has found many applications including adsorption, and various cabamation processes have been further achieved using different combinations of urea/mineral acids/mineral salts and other carbamate reagents (Khalil et al. 2002; Sirviö & Heiskanen 2019).

In the inorganic starch esters, substitution reaction creates linkages analogous to the ester linkage in the organic starch esters (Figure 9). Starch sulfates, starch phosphates and starch xanthates are important examples of inorganic starch esters. Starch sulfate synthesis and optimization (Figure 10) is thriving (Cui et al. 2009; Lin et al. 2009). They are used in paper and textile sizing, and are currently taking the stage among polysaccharide sulfates being studied for biological activities, such as high anti-virus activity, anti-HIV activity and anti-hemoglutination properties (Cui et al. 2009). Starch phosphates are synthesized by phosphorylation reactions with reagents such as; disodium phosphate (Figure 11), phosphorous oxychloride, sodium trimetaphosphate and sodium tripolyphosphate. The type of reagent and the reaction conditions determine the formation of mono- (substituted) and/or di- (cross-linked) starch phosphates (Chowdary et al. 2011; Sukhija et al. 2016; Rożnowski et al. 2017; Irawadi et al. 2018).

Starch xanthate is synthesized when starch is esterified with carbon disulfide in the presence of NaOH to obtain sodium dithiocarbonate ester of starch (Figure 12) (Krishnan & Attia 1987; Wang et al. 2009; Li et al. 2020). The dithiocarbonate groups (-OCSS^-) in starch xanthate are potential sites for adsorption of heavy metal ions.

In the dual modified SDAs, the selected esterification procedure is expected to introduce functional groups (i.e. the carbonyl/carboxylate at the ester linkages and others on the bulky substituents), capable of initiating or enhancing adsorption by interacting with AHMIs. Esterification using one or combinations of some di- or multi-functional reagents often results in the stepwise formation of mono-starch and then di-starch ester derivatives in which starch macromolecules are linked together. In these, the dual cross-linking and esterification is achieved using the esterifying reagent(s). Examples are some acetylation (Figure 17) and phosphorylation to di-starch (Figure 11(a) and 11(b)).

Etherification of starch occurs with reagents that react to substitute the hydroxyl groups on starch molecules and create ether linkages (St-OR) with the substituent group (Vandenbossche et al. 2015; Salah et al. 2019). Etherification reaction is sub-categorized into non-ionic or alkyl ether (with non-ionic hydroxypropyl or hydroxyethyl substituents), anionic (with carboxymethyl or sulfonic substituents), cationic (with ammonium, amino, imino or phosphonium substituents) and amphoteric (with both anionic and cationic substituents) starch etherification reactions (Figure 13) (Chen et al. 2015; Haroon et al. 2016). The reactions typically require an alkaline catalyst (initiator or activator) to initiate the chemical substitutions by creating nucleophilic (St-O⁻) sites at the O-H bonds; sodium hydroxide is commonly used (Figures 14 and 15) (Haroon et al. 2016; Masina et al. 2017; Tian et al. 2018). Carboxymethylation, hydroxypropylation, hydroxyethylation and some substitution reactions with monomeric reagents such as acrylamide and acrylic acid are important examples of the starch etherification reaction (Zhu et al. 2009; Hebeish et al. 2013; Chen et al. 2015; Masina et al. 2017). Starch ethers can be applied in areas such as coating, flocculants, drug delivery, additives, paper making and water–oil emulsion (Khalil & Aly 2002; Schmitz et al. 2006; Chen et al. 2015).

Carboxymethylation of starch is a simple, rapid and stepwise reaction. The activation step occurs with the alkaline catalyst (Figure 14, Step 1), while the carboxymethyl substitution step occurs with monochloroacetic acid (MCA) or sodium monochloroacetate (SMCA) (Figure 14, Step 2). The reaction is typically carried out in aqueous-organic solvents media to maintain optimum structural/granular integrity (Lawal et al. 2008; Spychaj et al. 2013; Akinterinwa et al. 2014; Masina et al. 2017). Degree of substitution in the reaction depends on the nature of starch and solvent, amount of the carboxymethylating agent, reaction time and temperature (Lawal et al. 2008; Nwokocha & Ogunmola 2008; Spychaj et al. 2013; Akinterinwa et al. 2014; Masina et al. 2017). Carboxymethyl starches (CMS) are used in various applications (Spychaj et al. 2013; Akinterinwa et al. 2014). However, their application in adsorption is based on enhanced swelling and the introduction of carbonyl (C=O) and carboxylate (COO^-) groups with potential affinity for heavy metal ions via chelation and ion-exchange mechanisms (Kim & Lim 1999; Palacios et al. 2004; Chauhan et al. 2010; Salah et al. 2019).

In hydroxypropylation and hydroxyethylation reactions, hydroxyl groups are substituted by hydroxypropyl and hydroxyethyl groups respectively. Propylene and ethylene oxides with NaOH (initiator) are typically used in the synthesis of hydroxypropyl (HPS) and hydroxyethyl (HES) starches respectively (Figure 15); however, other procedures have been successfully used (Schmitz et al. 2006; Lawal 2009; Masina et al. 2017; Fu et al. 2019). Compared to CMS, swellability and solubility are high in HPS and even higher in HES. This is due to the bulkiness of the substituents (CM ˂ HP ˂ HE), which will disrupt the inter- and intra-molecular forces and hydrogen bonds within the starch molecules (Ju et al. 2012; Masina et al. 2017). Hydroxypropyl sago starch was studied among other modified starches, in the removal of hardness (CaCO₃) from water (Racho & Namseethan 2017).

The reaction of starch with such monomeric reagents as acrylic acid results in the synthesis of starch copolymer ethers. In many cases, the long chain of the polymerized substituent is grafted on the starch molecule with the ether linkage (Figure 16) (Zhu et al. 2009; Hebeish et al. 2013).

Cross-linked-esterified starch derivatives are starch derivatives that are subjected to esterification reaction before or after the cross-linking reaction. This technique has been extensively utilized in the development of aqueous heavy metal ion adsorbent, and this will be discussed herein.

Semi- (esterified) and di-starch (cross-linked-esterified) derivatives of corn starch, i.e. SI and DI respectively, were synthesized using different concentrations (0.56 and 0.73 M respectively) of itaconic acid and used in the adsorption of Pb²⁺, Cu²⁺, Ni²⁺ and Zn²⁺ (Soto et al. 2016). The reaction (Figure 17) showed the interaction of itaconic acid with starch macromolecules at different stages of the reactions. Degree of substitution was slightly higher in SI than DI (0.0032 and 0.0030 respectively), and this was attributed to the higher cross-linking in DI. Water uptake (Figure 18) and some other characteristics of the derivatives show adequate substitution and cross-linking degrees that made them suitable for their use in heavy metal adsorption. Results of heavy metal adsorption using SI, DI and the native corn starch is presented in Table 1. Esterification (in SI) increased the adsorption of both Pb²⁺ and Zn²⁺ compared to the native starch, while cross-linking (in DI) further increased the adsorption of Zn²⁺. Since the report did not prove that there was no cross-linking in SI, it can be assumed that the degree of cross-linking in SI is at the optimum level for the adsorption of Pb²⁺. Adsorption of heavy metals on these derivatives was attributed to porosity. However, the reduction in Cd and Ni adsorption after modification can also be attributed to selective steric and conformation effects of the substituent group.

Racho and Namseethan compared the efficiencies of mono- and di- (cross-linked) sago starch phosphates, among others, in the removal of hardness (CaCO₃) from water (Racho & Namseethan 2017). Hardness removal was attributed to coordination reaction between Ca²⁺ and the phosphate groups, i.e. Ca(R-PO₃). The hardness removal efficiency was higher with mono-starch phosphate (53.41%) than di-starch phosphate (10.50%); however, turbidity, chemical oxygen demand (COD) and total dissolved solids (TDS) increased more in the water treated with mono-starch than di-starch phosphate (Racho & Namseethan 2017). This result showed that the mono-starch phosphate derivative requires optimum cross-linking to stabilize its structure and minimize its solubility/dispersability, which is causing the increase in turbidity, COD and TDS, whereas the di-starch phosphate requires optimum substitution with phosphate groups (i.e. mono-starch phosphates with less steric hindrance), to improve mass transfer and coordination.

Insoluble starch xanthate (ISX) is a derivative that is chemically cross-linked to make it insoluble in water and then xanthated. It was originally developed at the US Department of Agriculture, Northern Regional Research Center, Peoria, Illinois, and it is one of the earliest SDAs used in removing heavy metals from wastewater (Wing 1983). Over the years, the synthesis of ISX has been modified to obtain derivatives with higher adsorption efficiency. Corn starch was cross-linked with ECH and esterified while optimizing CS₂ and NaOH dosage as well as the reaction temperature, to obtain ISX with optimum sulfur content and copper ion adsorption capacity (Li et al. 2020). Sulfur content of the ISX was 7.54% at optimum reaction conditions (NaOH: 16 mL, CS₂: 2.5 mL and reaction temperature 35 °C), and characterization showed applicable changes in structural and functional properties of the derivative compare to the corn starch. ISX was used in the removal of Cu²⁺ from 1 L of 25 ppm aqueous solution, and the maximum removal capacity recorded was 90.5% (Li et al. 2020). Mohammed and Hendriks showed that the sulfur content of ECH cross-linked starch xanthate derivatives indicates the amount of xanthate bonded to the starch molecules, and the ability of the derivative to chemically-interact with the heavy metals ions to form the insoluble complex (Mohammed & Hendriks 2017). The results obtained when starch xanthates with different sulfur contents were used in the adsorption of Pb, Cd and Cu are shown in Table 2.

The xanthate or dithiocabamate group is relatively not such a bulky substituent. When substituted directly on cross-linked starch, its adsorption potential may be limited by steric effects and close packing of macromolecules, which may hinder chelation and mass transfer. Optimization of this process has therefore been studied in another trend of work in which the dithiocarbamate group is introduced on another substituent on the starch molecule. This is expected to reduce close packing and steric effects, and improve swelling and chemical functionality of the adsorbent. In a stepwise process, Li and co-workers cross-linked corn starch with ECH, treated it with ethylenediamine to obtain cross-linked amino-starch (CAS), and then esterified CAS with CS₂ to obtain a starch xanthate derivative: dithiocarbamate starch (DTCS) (Li et al. 2004). CAS and DTCS were regarded as chelating starch derivatives, and were both used in adsorption of Cu²⁺. An endothermic process that follows pseudo-first order kinetics and the Freundlich isotherm was reported for both derivatives. Equilibrium experiments showed that DTCS exhibits higher adsorption capacity and faster regeneration than CAS (Li et al. 2004). Cheng and co-workers, in their three-step process (Figure 19), cross-linked starch with glycidyl methacrylate (GS), treated it with ethylenediamine (EGS) and finally xanthated it with NaOH/CS₂ (DMGS) (Cheng et al. 2013). At optimum conditions, the adsorption capacity of DMGS for the metal ions studies followed the order Cu²⁺ > Cd²⁺ > Co²⁺ > Zn²⁺ > Ni²⁺ > Mn²⁺. The experiment data for all the ions studied fitted pseudo-second order kinetics and the Langmuir isotherm model, thereby indicating chemisorption mechanisms. The Fourier transform infrared (FTIR) spectra of DMGS before and after adsorption of the heavy metal ions are presented (Figure 20). The appearance of a new peak at 1,106 cm⁻¹ was attributed to coordination of sulfur atoms, hence showing that the dithiocarbamate groups chelate the metal ions (Cheng et al. 2013).

Heinze and co-workers synthesized starch phosphate carbamates by reacting wheat starch with phosphoric acid and urea (Heinze et al. 2003). The reaction was tuned, and it was observed that, as the amount of the modifying agents, reaction temperature and time were increased, the degree of phosphate group substitution also increased. Water retention value decreased and Cu²⁺ adsorption increased as the DS of phosphate group increased, and this was attributed to increased formation of di-starch phosphate cross-links in the derivatives (Heinze et al. 2003). Guo and co-workers cross-linked corn starch with ECH, before reacting with phosphoric acid and urea to obtain cross-linked phosphate carbamate derivatives used in the adsorption of Pb²⁺ (Guo et al. 2006). The phosphate and carbamate group contents of the optimized cross-linked starch phosphate carbamate (CSPS) were recorded as 3.10 and 1.40 mmol/g respectively. The derivative reached adsorption equilibrium in 20 min, the equilibrium data fitted Langmuir isotherm model, and the process was endothermic. Maximum Pb²⁺ adsorption of 2.01 mmol/g was obtained in the work; however, this decreased to 1.47 mmol/g after three cycles of reuse (Guo et al. 2006). Using same methods as Guo et al. (2006), CSPS was synthesized and used in the adsorption of Zn²⁺ (Guo et al. 2009). Adsorption equilibrium was reached in 40 min and the maximum adsorption was 2.0 mmol/g. Adsorption data fitted the pseudo-second order kinetic model and Langmuir isotherm model, and the process was endothermic (Guo et al. 2009).

Cross-linked carboxymethyl starches (CCMS) are the most common cross-linked-ether starch derivatives used as adsorbents, and this has been attributed to optimum swellability, porosity, chelation and ion-exchange on the carbonyl/carboxylate groups of the carboxymethyl substituents (Figures 21 and 25), as well as other functional groups from the cross-linking reagent (Kim & Lim 1999; Chen et al. 2012; Wang et al. 2012; Musarurwa & Tavengwa 2020).

Kim and Lim cross-linked corn starch with POCl₃, carboxymethylated it with different amounts of SMCA in ethanol solution, and studied the derivatives with the removal of Cu²⁺, Pb²⁺, Cd²⁺ and Hg²⁺ (Kim & Lim 1999). Adsorption capacity was shown to depend on the DS of carboxymethyl group (0.02–0.08) and the derivatives' pH (i.e. ionization of the CM), dosage in aqueous solution and solubility. The adsorption process was attributed to sodium ion-exchange on the carboxylate group, physical entrapment in the polymer matrix, and chelation (Kim & Lim 1999). Adsorption equilibrium was reached in 10 min; however, the adsorption capacity of the derivatives varied with the heavy metal ions studied (Table 3).

Banana starch was modified to CCSA and CCSB using POCl₃ and a mixture of STMP/STPP as cross-linkers respectively, while SMCA was used as the carboxymethylating reagent (Carmona-Garcia et al. 2009). Characterization of the derivatives confirmed successful modifications, while the chemical composition showed that phosphates content (CCSB: 0.2, CCSA: 0.01) and DS (CCSB: 0.09, CCSA: 0.06) were higher in CCSB than CCSA. The derivatives were used in the adsorption of Cu²⁺, Pb²⁺, Cd²⁺ and Hg²⁺, and the results obtained are as shown in Figure 22 (Carmona-Garcia et al. 2009).

Wang and co-workers synthesized CCMS by slightly cross-linking corn starch with ECH and carboxymethylating with monochloroacetic acid in a two-step alkali addition process (Figure 24) (Wang et al. 2012). Higher DS was recorded for the derivatives obtained in this study (0.43–0.59) due to low degree of cross-linking compared to Kim & Lim (1999). It was shown that even though the adsorption of heavy metal ion on the derivatives increased with DS due to chelation on the carboxylate group (Figure 25), ultrafiltration (i.e. ion trapping in polymer matrix) does not play any significant role in the adsorption process (Wang et al. 2012). Reusability of the derivatives was not studied; therefore the effect of solubility was not clearly shown. The low degree of cross-linking in the derivatives relegated ion entrapment in polymer matrix (ultrafiltration); this may also increase the solubility of the derivatives and limit their reusability. However, the high DS of the carboxymethyl groups in the derivatives will increase chelation with aqueous metal ions at fresh contacts. The chelation of heavy metal ions on the derivatives also depends on solution pH; for ions studied, it followed the order Pb²⁺ ˃ Cu²⁺ ˃ Cd²⁺ ˃ Zn²⁺ ˃ and Ni²⁺ (Wang et al. 2012).

Chen and co-workers later synthesized CCMS with a DS of 0.10 from corn starch using ECH and SMCA, and studied the equilibrium, kinetics, isotherms and thermodynamics of Pb²⁺ and Cd²⁺ adsorption on the derivative (Chen et al. 2012). Equilibrium studies showed that Pb²⁺ is better adsorbed than Cd²⁺ with maximum adsorption of 80.0 and 47.0 mg/g respectively. However, the adsorption capacities depend on the initial metal-ion concentration, contact time, solution pH, and temperature. The pseudo-second order (signalling chemical adsorption) and partly intra-particle diffusion were revealed as the rate-determining steps in the adsorption process of both ions. The adsorption of both ions was feasible: it followed Langmuir and Dubinin–Radushkevich isotherms with parameters showing that ion-exchange may be involved in the process. Thermodynamic parameters of the process showed that the adsorption of both ions was spontaneous and exothermic (Chen et al. 2012). Basri and co-workers carboxymethylated sago starch with SMCA and cross-linked it with lactic acid under irradiation to synthesize CMSS-acid hydrogel used in adsorption of Pb²⁺, Cu²⁺ and Cd²⁺ (Basri et al. 2016). Successful substitution and cross-linking reactions were confirmed from the characterization of the derivative. Adsorption equilibrium studies for the metal ions showed that equilibrium was reached within 60 min in 40 ppm aqueous solutions at pH 4, and 30 °C. At optimum condition, 0.050, 0.150 and 0.200 g of CMSS-acid hydrogel achieved 94.0%, 87.8% and 84.4% uptake of Pb²⁺, Cu²⁺ and Cd²⁺ respectively. The adsorption process was exothermic and kinetic parameters fitted the pseudo-second order equation for the three ions; however, the isotherm study showed that the adsorption of Pb²⁺ fitted the Freundlich model, while Cu²⁺ and Cd²⁺ fitted the Langmuir model (Basri et al. 2016). Pant and co-workers carboxymethylated corn starch with SMCA and cross-linked the derivative by irradiating it with different doses (1, 2, 3, 5, 10, 20, and 50 kGy) of electron beam (Pant et al. 2010). The highest degree of cross-linking was obtained in a 50% (w/w) aqueous CMS solution irradiated at 2 kGy electron beam dosage. The cross-linked carboxymethylated derivative was used in the removal of Cu²⁺ and Cd²⁺ from aqueous solutions. Both pH of metal solution and degree of cross-linking (gel content) of CCMS affected the metal uptake capacity. The maximum capacity for Cu (62%) and Cd (46.4%) was achieved at pH 5 and with the derivative of the highest cross-linking density. Removal of heavy metal ions was characterized and attributed to both chelation on carboxymethyl groups and physical entrapment of the ions in the cross-linked polymer network (Pant et al. 2010).

Some derivatives have been synthesized by subjecting the cross-linked starch to more than one etherification reactions. Xu and co-workers cross-linked corn starch with ECH, etherified it with 3-chloro-2-hydroxypropyltrimethylammonium chloride and then with MCA (carboxymethylation) to obtain cross-linked amphoteric starches (Figure 26) (Xu et al. 2003, 2004, 2005). The derivatives were studied for the adsorption of Cr⁶⁺, Cu²⁺ and Pb²⁺, and the process depended on DS, pH, adsorbent dose and initial concentrations. The effect of DS and pH on Cr⁶⁺ however contradicts that of Cu²⁺ and Pb²⁺. The adsorption of Cr⁶⁺ (in Cr₂O₄^-2 form) decreased while that of Cu²⁺ and Pb²⁺ increased as the DS of carboxymethyl group (anionic group) was increased in the derivatives. This is because the adsorption of dichromate ion is due to electrostatic trapping by cationic quaternary ammonium group, which is hindered by the anionic carboxylate group (-COO^-), and vice versa for Cu²⁺ and Pb²⁺. Also, adsorption of dichromate ions increased as pH decreased towards the acidic medium (below 4); this was attributed to the protonation of the carboxylate group (to –COOH), which reduced its repulsive effects on the dichromate ion. In contrast, Cu²⁺ and Pb²⁺ are attracted to the anionic group, hence its protonation as pH decreased reduced their adsorption. Adsorption of Cr⁶⁺ is exothermic while that of Cu²⁺ and Pb²⁺ is endothermic. Experimental data also showed that the adsorption of Cr⁶⁺, Cu²⁺ and Pb²⁺ fitted Langmuir, Freundlich and Langmuir isotherm models respectively, which showed that the adsorption process of the ions consists of different mechanisms (Xu et al. 2003, 2004, 2005).

Sancey and co-workers also synthesized cross-linked carboxymethyl amphoteric starch derivative (Sancey et al. 2011). In their work starch was cross-linked with 1,4-butanediol diglycidylether before the stepwise etherification with 2,3-epoxy-propyltrimethylammonium (in the presence of NH₄OH), and then with MCA (carboxymethylation) (Sancey et al. 2011). The highly functionalized and stabilized product obtained was studied for the remediation of industrial discharges. The removal of pollutants was efficient; however, it depended on the adsorbent dosage and contact time. Adsorption equilibrium for all the heavy metal ions studied (Zn, Pb, Cu, Ni, Fe and Cd) was reached within 60 min, and they were all removed beyond permitted limits. Turbidity and consequently COD was reduced in the wastewater, as well as other inorganic and organic matters causing phytotoxicity (Sancey et al. 2011).

Hood and O'Shea showed that adsorption of Ca²⁺ improved on a cross-linked hydroxypropyl derivative (hydroxypropylated di-starch phosphate: HDP) of tapioca starch. Adsorption of Ca²⁺ on both modified and unmodified starch reached equilibrium within 20 min, and depends on the initial concentration and temperature of the aqueous solution. With pH changed from 3.4 to 6.3, adsorption capacity of the HDP increased while the unmodified starch was unaffected, showing that the process is ionic/non-ionic on HDP but non-ionic on unmodified starch. When gelatinized, the adsorption capacities of both starches were eliminated, showing that the process also depends on a granular structure. Higher adsorption capacity was obtained with the HDP (86.5 μg/g) than the unmodified starch (64.0 μg/g) (Hood & O'Shea 1977). An intermediate 3-chloro-2-hydroxypropyl cross-linked starch (CHCS) was prepared by treating ECH cross-linked starch (CS) with more ECH in the presence of HClO₄, followed by the amination of the ether substituent to produce ethylenediamine modified cross-linked starch (CAS) (Figure 27), used in the removal of Cr⁶⁺ from aqueous solution (Cheng et al. 2009). Adsorption was most favorable at pH 4, reached equilibrium within 4 h, and achieved maximum capacity of 91% Cr⁶⁺ removal from 14 ppm solution. Experimental data fitted pseudo-second order kinetics and Freundlich isotherm model, and the adsorption of Cr⁶⁺ was attributed to electrostatic attraction between −NH⁺₃···HCrO₄⁻ and −NH⁺₂···HCrO₄⁻. This was further confirmed by splits, shifts and appearance of new peaks on the FTIR spectra of Cr-bound CAS (CAS-Cr) compared to CAS (Figure 28).

Klimaviciute and co-workers carried out cross-linking and cationic etherification (using NaOH as catalyst) on potato, corn and wheat type-A and type-B starches to obtain CCSP, CCSC, CCSWA, and CCSWB respectively, using ECH and 2,3-epoxypropyltrimethylammonium chloride as the cross-linking and etherifying reagents respectively (Klimaviciute et al. 2010). The derivatives were used in the adsorption of Cr⁶⁺ in dichromate form (Cr₂O₇⁻²), and the processes were feasible and efficient. DS of the cationic group on the derivatives increased with the granular size of the native starches; this also increased ion-exchange and hence the adsorption efficiencies of the derivatives. Langmuir, Dubinin–Radushkevich and Temkin models indicated the best fitting to the experimental data, and adsorption favorability followed the order CCSWB < CCSWA ≈ CCSC < CCSP. Furthermore in the study, an organic (benzyltrimethylammonium hydroxide) catalyst was used in the cationic etherification of potato starch to obtain CCSPO. X-ray diffraction spectra showed that the organic catalyst increased the amorphous degree compared to CCSP, which resulted in an increase in dichromate removal. This may therefore serve as a route to obtain efficient adsorbents from native starches with smaller granule (Klimaviciute et al. 2010).

In another trend, the cross-linked starch is subjected to both esterification and etherification reactions to improve its stability and enrich its functionality as a heavy metal adsorbent. Xing and co-workers cross-linked corn starch with ECH and etherified it with 2,3-epoxypropyltrimethylammonium chloride followed by esterification using maleic anhydride to obtain cross-linked cationic starch maleates (CCSM) (Figure 29), used in the adsorption of Cr⁶⁺ (Xing et al. 2006). The adsorption of Cr⁶⁺ decreased (from 35.71 to 27.10 mg/g) as the DS of maleic group increased (from 0.02 to 0.07) in the CCSM derivatives. This was attributed to prevention of electrostatic attraction between Cr⁶⁺ and the quaternary ammonium group in the CCSM by the anionic maleic group. The adsorption process was further shown to depend on adsorbent dosage, solution pH and temperature. The process was exothermic and the data fitted Langmuir isotherm model (Xing et al. 2006).

Xiang and co-workers also developed DTCS and two other derivatives: dithiocarbamate enzymolysis starch (DTCES) and dithiocarbamate mesoporous starch (DTCMS), and used them in the adsorption of some heavy metals (Xiang et al. 2016). Adsorption capacities of the derivatives follow the order: DTCMS ˃ DTCES ˃ DTCS, while the adsorption of heavy metal ions on the derivatives follow the order: Cu²⁺ > Ni²⁺ > Cr⁶⁺ > Zn²⁺ > Pb²⁺. Adsorption was shown to be hindered in acidic solutions due to the adsorbent's preferential interaction with hydrogen ions. Kinetics of the adsorption processes (of Cu²⁺ and Pb²⁺) fitted the pseudo-second order model, and chelation with the xanthate group was assumed to be involved in the adsorption mechanism (Figure 30) (Xiang et al. 2016).

For flocculation and adsorption of Cu²⁺, synergetic effect of cross-linked starch xanthate (CSX) and cross-linked starch-graft-polyacrylamide (CSA) was studied in cross-linked starch-graft-polyacrylamide-co-sodium xanthate (CSAX) synthesized by graft copolymerization reaction of corn starch, acrylamide, sodium xanthate and ECH as the cross-linking reagent (Chang et al. 2008). Removal of Cu²⁺ and turbidity on the derivatives depended on the adsorbent dosage; however, Cu²⁺ removal effectiveness followed CSAX ˃ CSX ˃ CSA, while turbidity removal effectiveness followed CSA ˃ CSAX ˃ CSX (Figure 31(a) and 31(b)) (Chang et al. 2008).

Feng and Wen synthesized N,N′-methylenebisacrylamide CSX and grafted with sodium acrylate and acrylamide in an etherification reaction to obtain a product studied for the adsorption of Pb²⁺ and Cd²⁺ (Feng & Wen 2017). Effects of the reaction parameters such as; starch xanthate to monomer (sodium acrylate and acrylamide) ratio and initiator (potassium persulfate) and cross-linker dosage on the adsorption of Pb²⁺ and Cd²⁺ were studied as shown in Figure 32(a)–32(c) respectively. Adsorption of Pb²⁺ and Cd²⁺ on the derivative was hindered in acidic solutions, both followed Freundlich isotherm, and the highest quantities of Pb²⁺ and Cd²⁺ adsorption were 47.11 mg/g and 36.55 mg/g, respectively (Feng & Wen 2017).

Cross-linking, esterification and/or etherification reactions have been involved in some stepwise procedures followed to improve various limiting botanical properties of native starches and synthesize derivatives optimized for applications as AHMI adsorbents. Adsorption of AHMIs on SDAs was attributed to chelation, ion-exchange, electrostatic interactions and a physical trapping polymer matrix. To enhance these mechanisms, different cross-linking reagents are used in the cross-linking reactions at different degree of cross-linking to optimize the porosity and structural stability of the starch polymer network and achieve an efficient mass flow and AHMI trapping. Esterification and etherification reactions with selected reagents substitute some heavy metal interactive chemical species onto starch molecules via stable ester and ether linkages. The nature and DS of the substituent groups on SDA determine the AHMI selectivity and extent of binding or non-binding due to steric hindrances. Therefore, selection of reagent (hence substituent groups) and experimental control of the DS allowed optimization of AHMI binding/adsorption. SDAs have been studied for the rapid and efficient removal of divalent and hexavalent AHMIs; however, the efficiency of these products is still guided within a narrow range of process conditions such as; pH, temperature and initial concentration of aqueous solutions. Future SDAs from cross-link-esterification/esterification methods should be synthesized using novel modifying agents (such as higher molecular weight cross-linkers), and starches from novel botanical sources. Also, studies on the synergetic effects of SDAs and their composites with other adsorbents may form another path towards economical production of adsorbents with sustained efficiency within a wide stretch of process conditions.