The adsorption of Cd and Zn to alumina nanoparticles could be affected by numerous factors, including the presence of naturally occurring ligands, which are ubiquitous in the soil and water environment and in various wastewaters. They include low-molecular-weight organic acids such as citric and oxalic acid, high-molecular-weight acids such as humic acid (HA), and important inorganic ligands such as phosphate, sulfate, and carbonate. Previous research on bulk-sized minerals showed that varying inorganic and organic ligands may enhance Cd or Zn adsorption to goethite or kaolinite by forming metal–ligand–surface ternary complexes, by increasing the surface electrostatic potential, or by formation of metal-ligand precipitates (Davis and Bhatnagar, 1995; Collins et al., 1999; Stietiya et al., 2011). On the other hand, Pb adsorption in soil decreased when the formation of Pb–mineral surface complexes in the presence of citrate was unable to compete with stable Pb–citrate complexes in solution (Schwab et al., 2005). Furthermore, citrate enhanced Cd adsorption to goethite and montmorillonite when present at low concentrations but inhibited it at high concentrations (Lackovic et al., 2004; Huang et al., 2010). Phosphate enhanced Zn or Cd adsorption to iron oxide surfaces due to electrostatic interaction (Diaz-Barrientos et al., 1990; Collins et al., 1999; Wang and Xing, 2004), whereas it was found to reduce Cd adsorption to hematite and soils due to formation of Cd–P complexes in solution (Krishnamurti et al., 1999; Li et al., 2006). Clearly, different results have been observed regarding the impact of ligands on metal adsorption by minerals and soils. In addition, although previous studies were primarily focused on metal adsorption by bulk-sized minerals, few studies have investigated ligand effects on nanoparticle mineral surfaces. Limited studies have reported reduced Cu adsorption onto hydroxyapatite nanoparticles in the presence of low-molecular-weight organic acids (Wang et al., 2009). On the other hand, chemical or physical surface modification of Al2O3 nanoparticles with functional groups containing donor atoms such as oxygen, nitrogen, and sulfur enhanced adsorption of various metals (Savage and Diallo, 2005; Afkhami et al., 2010). It is imperative to understand the conditions under which adsorption of heavy metals is enhanced in the presence of ligands to optimize the adsorption efficiency of Al2O3 nanoparticles for various heavy metals. To the best of our knowledge, this is the first investigation that compares the impact of PO4, citrate, and HA, three common inorganic and organic ligands, on the adsorption of Cd and Zn to Al2O3 nanoparticles at various ligand:metal ratios. The objective of the study was to investigate Zn and Cd adsorption to Al2O3 nanoparticles as influenced by varying concentrations of PO4, citrate, and HA in mono-metal and binary-metal systems at pH 6.5. In support of the primary objective, the adsorption of PO4, citrate, and HA to the nanoparticles was characterized in the absence and presence of Cd and Zn.
Materials and Methods
Materials and Reagents
Aluminum oxide nanopowder (γ-Al2O3) was obtained from Sigma-Aldrich (catalogue #544833). The main characteristics of this product as provided by the supplier are: particle size <50 nm, melting point 2040°C, and density 3.97 g cm−3. The BET surface area was measured as 185 m2 g−1. Zinc, Cd, PO4, and citrate were prepared from Zn(NO3)2, Cd(NO3)2, NaH2PO4, and trisodium citrate dihydrate, respectively. Humic acid was obtained from Sigma-Aldrich (H16752) and was purified by acid washing and subsequent removal of ash content according to the International Humic Substance Society method (Swift, 1996). Humic acid was then dialyzed until free of Cl− (as tested using AgNO3), freeze dried, and homogenized before use (Stietiya et al., 2011). All reagents used in the present investigation were of analytical reagent grade. Ultra-pure water with resistivity of 18.2 MΩ was used throughout experimentation and was obtained from a Milli-Q Water System (Millipore Corp.).
Adsorption experiments were conducted to obtain the equilibrium isotherms for Cd and Zn. Three Zn and Cd adsorption systems were constructed: Zn mono-metal (adsorption of Zn only), Cd mono-metal (adsorption of Cd only), and Zn/Cd binary-metal (adsorption of Zn and Cd) systems. The initial Zn and Cd concentrations ranged from 0.05 to 1 mmol L−1 as nitrate salts at an adsorbent dosage level of 1 g L−1 or as a solid:solution ratio of 1:1000. In the binary-metal system, the concentration of Zn was identical to that of Cd at each initial concentration. The background electrolyte concentration for the systems was 10 mmol L−1 NaNO3 to dominate ionic strength, and pH was fixed at 6.5 using 0.1 mol L−1 HNO3 or 0.1 mol L−1 NaOH. After 1 wk of equilibration, the samples were centrifuged, and equilibrium concentrations of Cd and Zn were determined using ICP–AES (Spectro Analytical Instruments). The equilibrium data obtained were then fit to the Langmuir and Freundlich adsorption isotherms. The Langmuir model is expressed as:where qe is amount of Zn or Cd adsorbed per unit weight of Al2O3 (mmol kg−1), Ce is the equilibrium concentration (mmol L−1), Q0 is the monolayer adsorption capacity (mmol kg−1), and b is the constant related to the free energy of adsorption (L mmol−1). The Freundlich model is expressed as:where KF is the constant indicative of the relative adsorption capacity of Al2O3 nanoparticles (mmol kg−1), and N is the constant indicative of the intensity of adsorption. Isotherms were also constructed for citrate, HA, and PO4 in the absence of Cd or Zn, in the presence of 1 mmol L−1 Cd, in the presence of 1 mmol L−1 Zn, and in the presence of 1 mmol L−1 Zn and Cd. Experiments were performed in triplicate.
Zinc and Cd adsorption to Al2O3 nanoparticles in the absence and presence of varying concentrations of PO4, citrate, and HA was studied. The impact of ligand concentrations on metal adsorption was studied in three metal systems: Zn mono-metal, Cd mono-metal, and Zn/Cd binary-metal systems. The initial concentrations of Zn and Cd in the systems were 1 mmol L−1 as nitrate salts, and ligand concentrations were 0, 0.25, 0.5, 1.0, and 2.0 mmol L−1. This resulted in ligand-to-metal molar ratios of 0:1, 0.25:1, 0.5:1, 1:1, and 2:1 in each of the systems investigated at an adsorbent dosage level of 1 g L−1 or a solid:solution ratio of 1:1000. Ligands were added to the systems 24 h before the addition of Zn or Cd metal solution to eliminate the possibility of metal-ligand precipitate formation, particularly with PO4. Citrate and PO4 concentrations were expressed in mmol L−1, and HA was expressed in mg L−1. The background electrolyte concentration was 10 mmol L−1 NaNO3, and pH was fixed at 6.5. The pH was continuously checked and adjusted using 0.1 mol L−1 HNO3 or NaOH. After 1 wk of equilibration, samples were centrifuged at 47,893 × g for 10 min, and the supernatant was filtered using 0.45-μm membrane filters. Equilibrium concentrations of Zn, Cd, and P were determined using ICP–AES (Spectro Analytical Instruments). A preliminary experiment confirmed the sufficiency of settling Al2O3 nanoparticles in 0.01 mol L−1 NaNO3 and found no Al (by ICP) in the supernatant after filtering with a 0.45-μm membrane filter. Equilibrium concentrations of citrate and HA were determined using a total organic carbon analyzer (Shimadzu). The amounts of Zn, Cd, citrate, HA, and PO4 adsorbed were calculated by the difference between the initial and equilibrium concentrations. Adsorption experiments were performed in triplicate.
Analysis of variance was performed using SPSS 17 (SPSS, Inc.), and comparison of means was undertaken using a Bonferroni test to determine any significant differences at P ≤ 0.05.
Results and Discussion
Zn and Cd Adsorption Isotherms
Figure 1 shows Zn and Cd adsorption isotherms in mono-metal and binary-metal systems. The equilibrium data for Zn and Cd adsorption gave generally satisfactory fits to the Freundlich and Langmuir models (R2 > 0.90), although Zn adsorption was a slightly better fit using the Freundlich model in both metal systems, and Cd adsorption gave slightly better fits to the Langmuir model (Table 1). The Freundlich model assumes that the uptake of Zn and Cd occurs on a heterogeneous Al2O3 nanoparticle surface, whereas the Langmuir model assumes uptake on a homogeneous surface by monolayer adsorption with no interaction between sorbed species (Sharma et al., 2009). Zinc adsorption was greater than Cd adsorption in both mono-metal and binary-metal systems (Fig. 1). The greater affinity of Al2O3 nanoparticles for Zn adsorption was likely attributed to chemical characteristics of ions, such as ionic radii, and hydrolysis constants (pKh). The smaller unhydrated ionic radius of Zn (0.074 nm) as compared with that of Cd (0.097 nm) suggests stronger electrostatic adsorption of the former (McBride, 1994; Antoniadis et al., 2007). The first hydrolysis products of Zn and Cd are 9.0 and 10.1, respectively. Metal ions having lower hydrolysis constants are characterized by higher electrostatic attraction to exchange sites (Bosso and Enzweiler, 2002; Yavuz et al., 2003; Yang et al., 2006; Unuabonah et al., 2007; Abollino et al., 2008; Srivastava et al., 2008; Huang et al., 2010). If chemisorption is the dominant adsorption mechanism, the more electronegative Cd ion would form stronger covalent bonds with O atoms of the Al2O3 surface than Zn and would hence be expected to adsorb preferentially over Zn (McBride, 1994; Mahdavi et al., 2012). Because this was not the case in our study, it is likely that adsorption to surface sites was mainly electrostatic in nature. On the other hand, we could not rule out the possible formation of inner-sphere surface complex (chemisorption) or surface precipitation because other research based on extended X-ray absorption fine structure spectroscopy showed that Zn sorption to bulk-sized (>100 nm) α-phase–dominated Al2O3 minerals formed inner-sphere surface complexes at low sorption densities and formed a precipitate phase at higher concentrations (Trainor et al., 2000). Previous studies also reported that Zn formed outer-sphere complexes with ferrihydrite irrespective of pH or surface loading (Trivedi et al., 2001).
|Metal ion||System||Freundlich isotherm
|L kg−1||mmol kg−1||L mmol−1|
Zinc and Cd adsorption rates were higher in their respective mono-metal systems than in the binary-metal system, especially at higher equilibrium concentrations (Fig. 1). The distribution coefficient (KF) for Zn adsorption in mono-metal system was 366.9 L kg−1, as compared with 241.1 L kg−1 in the binary-metal system (Table 1). Higher KF values indicate stronger adsorption and lower metal solubility (Antoniadis et al., 2007). The lower KF value in the binary-metal system indicates that Zn adsorption was suppressed in the presence of Cd (Unuabonah et al., 2007). The selectivity of Zn and Cd adsorption to Al2O3 nanoparticles based on KF values was in the order of Zn mono-metal > Zn binary-metal > Cd mono-metal > Cd binary-metal. Competition between Zn and Cd for the same binding sites was likely the cause of reduction in adsorption in the binary-metal systems. The adsorption rates of Zn and Cd were favorable as indicated by the Freundlich model, where values of N between 0 and 1 represent favorable adsorption (Antoniadis et al., 2007; Sheela et al., 2012) (Table 1).
Similar to Zn, Cd adsorption was higher in the mono-metal system than in the binary-metal system (Fig. 1). The value of adsorption capacity, or the maximum uptake for Cd in mono-metal system (Q0mon), was 73.4 mmol kg−1, as compared with 44.0 mmol kg−1 in the binary-metal system (Q0bin). The value of Q0bin/Q0mon being less than unity indicates the suppression of Cd adsorption in the presence of Zn (Adebowale et al., 2006; Unuabonah et al., 2007). The feasibility of fitting the Langmuir isotherm can be expressed using the separation factor, RL, defined as:where Co is the initial concentration (mmol L−1) of Cd and Zn in solution, and b is the Langmuir constant. The shape of the isotherms is unfavorable at RL > 1, favorable at 0 < RL < 1, or linear at RL = 1 (Hao et al., 2010; Sheela et al., 2012). The RL values for adsorption of Zn and Cd in mono-metal and binary-metal systems at all initial concentrations are <1 (Fig. 2), indicating the suitability of Al2O3 nanoparticles for the removal of Zn and Cd from solution (Doğan et al., 2000; Sheela et al., 2012).
The competitive effects of metal adsorption in bulk-sized minerals have been widely documented. For example, the presence of Cu was found to reduce Zn adsorption to goethite, whereas Pb was shown to reduce Cd adsorption to kaolinite (Juang and Chung, 2004; Adebowale et al., 2006). Adsorption of Cu, Ni, Cd, and Pb to nano-size particles of Fe3O4, ZnO, and CuO was also found to be lower in competitive-metal systems than in the respective mono-metal system of each metal (Mahdavi et al., 2012). Mahdavi et al. (2013) reported that Cd adsorption by Al2O3 nanoparticles in the competitive system of Cu, Ni, and Pb decreased by as much as 80% in comparison with Cd in the mono-metal system. Our results in this study also showed that Cd adsorption decreased by 60% from 73.4 to 44 mmol kg−1, whereas Zn adsorption decreased by 37% from 298.2 to 187.4 mmol kg−1 (Table 1). Although the magnitude of competition among metals may be different depending on metal and mineral types, the general phenomenon of competitive effect of metal adsorption in nano-sized mineral systems appears to be consistent with that observed in bulk-sized mineral systems.
The surface area of Al2O3 nanoparticles has been reported to range from 43 to 411 m2 g−1 depending on the method of preparation (Hua et al., 2012), which is generally larger than that of bulk-size Al2O3 particles (9.3–16 m2 g−1) (Chen et al., 1973; Boily and Fein, 1996; Trainor et al., 2000). As shown in this study, Al2O3 nanoparticles yielded 10 to 30 and 40 to 100 times higher adsorption capacities of Zn and Cd, respectively, in mono-metal systems as compared with those reported for bulk-sized Al2O3 (Hachiya et al., 1984; Benjamin and Leckie, 1980; Trainor et al., 2000). These Zn and Cd adsorption capacities of Al2O3 nanoparticles were also much greater (2–100 times) than those reported for bulk-sized gibbsite or bayerite (Kinniburgh et al., 1976; Shuman, 1973; Weerasooriya et al., 2002) as the surface of Al2O3 was known to generally hydroxylate in aqueous solution, forming gibbsite or bayerite-like surfaces (Laiti et al., 1998; Eng et al., 2000). The higher surface area of nanoparticles was likely the cause for the increased Zn and Cd adsorption capacity. Recently, Rahmani et al. (2010) reported a maximum Zn adsorption capacity of 899 mmol kg−1 for Al2O3 nanoparticles with an even higher surface area of 206 m2 g−1 and a smaller mean nanoparticle size of 7.21 nm. In this study, we did not examine the possibility of conglomeration of Al2O3 nanoparticles during the adsorption experiment, which could reduce the adsorption capacity of metals (Mahdavi et al., 2013). However, the experimental dosage of Al2O3 nanoparticles in this study was not high (i.e., it was less than the critical limit of dosage of <4 g L−1 for potential conglomeration) (Rahmani et al., 2010), and therefore the likelihood of this happening was low. Conglomeration of nanoparticles could be controlled using stabilizers such as carboxylic acids and polymers (Bahrami et al., 2012). Additionally, decreasing Al2O3 nanoparticle adsorbent dosage below critical limits reduces conglomeration (Mahdavi et al., 2013).
Impact of Ligands on Zn and Cd Adsorption
Figure 3 shows Zn and Cd adsorption to Al2O3 nanoparticles in the presence of increasing PO4, citrate, and HA in mono-metal and binary-metal systems where initial concentrations of Zn and Cd were at 1 mmol L−1. In the absence of any of the complexing ligands (control), the amount and trend of Zn and Cd adsorption in both the mono-metal and the binary-metal systems were consistent with those in isotherms (Fig. 1). The impact of ligands on adsorption of Zn and Cd by Al2O3 nanoparticles varied by ligand type, ligand concentration, and metal system (Fig. 3). In general, the presence of PO4 and HA enhanced Zn and Cd adsorption in all systems, whereas citrate had both inhibitory and beneficial effects on adsorption, depending on metal system and citrate concentration. Zinc reached near-complete adsorption (996 mmol kg−1) at 1.0 mmol L−1 PO4 in the mono-metal system and at 2.0 mmol L−1 PO4 in the binary-metal system. Cadmium also reached near-complete adsorption at 2 mmol L−1 PO4 (PO4:Zn ratio of 2:1) in the mono-metal and binary-metal systems (Fig. 3A). Although HA had similar effects to PO4 in enhancing metal adsorption, complete adsorption of Zn and Cd was not observed in any of the systems at these HA concentrations (Fig. 3C). With HA, the largest increase in adsorption was for Cd in the mono-metal system (from 5% [46 mmol kg−1] to 58% [583 mmol kg−1]), and the smallest increase in adsorption was for Zn in mono-metal system (from 48% [480 mmol kg−1] to 66% [661 mmol kg−1]) at HA concentrations ranging from 0 (control) to 240 mg L−1 HA. In the citrate systems where adsorption of Zn and Cd was enhanced, adsorption had not exceeded 33 mmol kg−1 (33%) of the total initial concentration of Zn (Fig. 3B). Thus, PO4 was the most effective ligand in comparison with HA and citrate for promoting Zn and Cd adsorption.
Previous studies on bulk-size minerals had shown enhanced adsorption of Zn, Cd, and other metals to mineral surfaces by PO4 due to ternary complex formation or due to increases in negative charge of the inner Helmholtz plane of the mineral surface as a result of specific adsorption of PO4 (Diaz-Barrientos et al., 1990; Collins et al., 1999; Wang and Xing, 2004; Stietiya et al., 2011; Ren et al., 2012) and HA (Davis and Leckie, 1978; Vermeer et al., 1999; Lai et al., 2002; Lai et al., 2002; Arias et al., 2002; Stietiya et al., 2011). Conversely, Cd adsorption on hematite and in soils was lower in the presence of PO4 likely due to the formation of soluble PO4 complexes like CdHPO4 or due to PO4 blocking Cd sorption sites (Krishnamurti et al., 1999; Li et al., 2006). In addition, PO4 was found to increase Cu adsorption capacity of bulk-size γ-Al2O3 from 52.7 to 129.6 mmol kg−1 at low concentration, but PO4 concentrations >2.4 mg L−1 P (or 0.03 mmol L−1 PO4) had little influence on Cu adsorption capacity (Ren et al., 2012). Humic acid, which binds directly on iron oxide surfaces, may block surface sites, thus decreasing metal adsorption (Zuyi et al., 2000; Lai et al., 2002). Fulvic acid did not affect Zn adsorption to bulk-size Al2O3 surface at pH <7 (Zuyi et al., 2000). The effect of humic substances on metal sorption appears to be dependent on the nature of the metal oxides and the humic substance (Zuyi et al., 2000). Nonetheless, our results showed strong PO4 enhancement of Zn and Cd adsorption with concentration up to 2 mmol L−1, indicating the significant impact of PO4 on Zn and Cd adsorption to Al2O3 nanoparticle surface (Fig. 3A). These results also confirmed the general positive impact of HA in enhancing Zn and Cd adsorption to Al2O3 nanoparticle surfaces at the studied solid to solution ratio and HA concentration ranges (Fig. 3C).
Contrary to PO4 and HA, the impact of citrate on Zn and Cd adsorption varied among the systems (Fig. 3B). Citrate had a positive impact on Zn adsorption in the binary-metal system and on Cd adsorption in both metal systems (Fig. 3B). In these systems, adsorption was enhanced specifically at low citrate concentrations between 0.25 and 1.0 mmol L−1 or at a citrate:metal ratio between 1 and 2. At the highest citrate concentration of 2.0 mmol L−1, adsorption of Zn in the binary-metal system and Cd in the mono-metal system was lower than at any other concentration. Enhancement of heavy metal adsorption at low organic acid concentration and suppression at high concentration has been reported for Al2O3 and goethite bulk-size minerals (Liao, 2006; Huang et al., 2010). For instance, Huang et al. (2010) reported that Cd2+ adsorption to goethite was enhanced at low citric acid concentrations (1.0 mmol L−1) and was suppressed at 1.0 to 3.0 mmol L−1. The current study showed similar behavior of citrate in Al2O3 nanoparticles for Zn in binary-metal system and for Cd in both systems. In the Zn mono-metal system, however, the presence of citrate reduced Zn adsorption at all levels (Fig. 3B). The results suggest that the trend of enhanced metal adsorption to bulk-size mineral surfaces at low citrate concentrations does not apply to Zn adsorption to Al2O3 nanoparticles when present alone in the system (Zn mono-metal system). This may be due to the weak adsorption of citrate by nano-sized Al2O3 and the formation of a strong solution citrate–Zn complex that prevents Zn from bonding to Al2O3 nanoparticle surfaces. On the other hand, the fact that Zn in the binary metal system was enhanced by citrate at the low concentration (<2 mmol L−1) could indicate that the surface-bonded Cd may further bind solution citrate-Zn species and could therefore enhance Zn adsorption in the binary metal system in the presence of citrate.
To better understand the effects of ligands on Zn and Cd adsorption to Al2O3 nanoparticles, adsorption isotherms of PO4, citrate, and HA in the mono-metal systems and binary-metal system are presented in Fig. 4. Adsorption of these ligands was generally better fit to the Langmuir model than to the Freundlich model in all systems except for HA in both metal systems, which did not fit due to too strong HA adsorption by Al2O3 nanoparticles (Table 2). The separation factor confirmed the favorable adsorption of PO4, citrate, and HA to Al2O3 nanoparticle surface (Fig. 5). The isotherms revealed that the presence of Zn or Cd enhanced the adsorption of PO4, citrate, and HA to the Al2O3 nanoparticle surface in these systems because the absence of both metals showed generally lower ligand adsorption (Fig. 4). The comparison of PO4 adsorption capacity (Q0) in different metal systems showed that the presence of both Zn and Cd enhanced adsorption of PO4, whereas the presence of Zn alone had the greatest adsorption of citrate and Zn or Cd alone had the greatest adsorption of HA (Fig. 4; Table 2). These results indicated that the enhanced adsorption of Zn and Cd in the metal systems (Fig. 3) was accompanied by generally enhanced adsorption of PO4, citrate, and HA (Fig. 4). Accordingly, the different impact of PO4 and HA on Zn and Cd adsorption from that of citrate may be explained by differences in the ligand adsorption (Fig. 4). Most of the PO4 and HA was adsorbed to Al2O3 nanoparticle surfaces, especially at low ligand concentrations, whereas large portions of citrate remained nonadsorbed, specifically at high concentrations. Considerable equilibrium concentrations of citrate indicate that Al2O3 surface sites were saturated with the ligand, which would typically allow excess citrate to compete with surface sites for Zn or Cd complexation (Boily and Fein, 1996; Liao, 2006). Citrate has three pKa values (3.13, 4.76, and 6.39) and would be in the dissociated Cit3− form at pH 6.5, which would provide high chelating ability for Zn and Cd as compared with the partially dissociated H2Cit− and HCit2− species (Qin et al., 2004; Johnson and Loeppert, 2006; Abollino et al., 2008). Therefore, the dominant Cd and Zn solution species in the presence of citrate are likely CdCit−1 and ZnCit−1, respectively (Boily and Fein, 1996). The same complexes with carboxyl groups would be assumed for HA because carboxyls account for more than 50% of total acidity and have pKa in the range of 4 to 6, whereas phenols range from 9 to 11 (Lai et al., 2002). Nevertheless, the impact of these two ligands on Zn and Cd adsorption was quite different, indicating the varying interaction with Al2O3 nanoparticle surface groups. Stietiya et al. (2011) also showed that citrate suppressed adsorption at pH >5.7, whereas HA and PO4 increased Zn adsorption to kaolinite.
|Metal ion||System||Freundlich isotherm
|L kg−1||mmol kg−1||L mmol−1|
|PO4||absence of metals||999.4||0.28||0.86||983.6||1.66||0.93|
|Citrate||absence of metals||157.2||0.21||0.82||171.0||11.8||0.89|
|Humic acid||absence of metals||–||–||–||1.13 × 105 mg kg−1||0.06||0.93|
The expected adsorption complexes formed in the absence of ligands were likely ≡Al-OCd+ and ≡Al-OZn+ despite the high pHPZC (8.9–9.1) of Al2O3 (Pacheco et al., 2006). Previous modeling attempts of coadsorption of Cd and citrate by α-Al2O3 suggested that the surface species ≡AlCitCd0 was likely dominant between pH 5 and 7.9, whereas ≡AlOCd+ was at pH >8.0 due to increasing Al surface deprotonation and reduced citrate adsorption at high pH (Boily and Fein, 1996). Although similar surface species could be expected in this study for Cd and Zn adsorption by γ-Al2O3 nanoparticles, the formation Cd- and Zn-citrate precipitates has been reported (Collins et al., 1999). The exact formation of surface complex species in the presence of these ligands would require confirmation of detailed solid speciation (Stietiya et al., 2011).
The fact that adsorption of both Zn and Cd was generally enhanced as a result of the presence of these ligands and vice versa suggests that several mechanisms may be at work. Although it is possible that adsorption of PO4, citrate, and HA reduced the overall positive charge of the nano-sized Al2O3 surface, the beneficial impact of these ligands on metal adsorption would indicate the following processes: (i) PO4, citrate, and HA formed soluble complexes with Cd and Zn, and such complexes had high affinity for an Al2O3 nanoparticle surface; (ii) ternary complex formation; or (iii) surface precipitation of HA with Zn and Cd as shown for bulk-size minerals (Collins et al., 1999; Arias et al., 2002; Ren et al., 2012). Regardless of the unconfirmed underlining mechanisms, our results showed enhancing Zn and Cd adsorption in mono- and binary-metal systems with increasing PO4 concentration up to 2 mmol L−1 or HA concentration up to 240 mg L−1. On the other hand, citrate only exhibited increased metal adsorption in Cd mono- or Zn and Cd binary systems with optimal concentration between 0.5 and 1.0 mmol L−1 but decreased Zn adsorption in a mono-metal system throughout the citrate concentrations studied.
This study demonstrated that Zn and Cd adsorption by nano-sized Al2O3 was generally much higher than the reported adsorption capacities of these metals by bulk-sized alumina. Zinc had higher affinity to the γ-Al2O3 nanoparticle surface than Cd. For both Zn and Cd, the adsorption in the binary-metal system was lower than in their respective mono-metal systems, indicating competition for the same surface sites of nano-Al2O3 minerals, a trend that was similar to that reported for bulk-sized Al2O3. The presence of PO4 and HA enhanced Zn and Cd adsorption in all systems, whereas citrate reduced Zn adsorption in the mono-metal system but increased adsorption in the other metal systems at low citrate concentrations (0.5–1 mmol L−1). Removal of Zn or Cd from the systems by nano-Al2O3 minerals was generally accompanied by enhanced removal of PO4 and HA, suggesting the formation of a ternary surface complex or metal-ligand precipitation. Overall, Al2O3 nanoparticles are suitable for use as a sorbent for removing Zn and Cd from solution systems, which can be significantly enhanced by the presence of PO4 or HA. Citrate, on the other hand, can also promote solution Cd and Zn removal, especially the former, by Al2O3 nanoparticles at low citrate concentrations or at citrate:Cd concentration ratios between 1 and 2.