Steel Slag and Autoclaved Aerated Concrete Grains as Low-Cost Adsorbents to Remove Cd 2+ and Pb 2+ in Wastewater: Effects of Mixing Proportions of Grains and Liquid-to-Solid Ratio

: This study investigated the applicability of industrial by-products such as steel slag (SS) and autoclaved aerated concrete (AAC) grains (<0.105, 0.105–2, 2–4.75 mm) as low-cost adsorbents for simultaneous removal of Cd 2+ and Pb 2+ in wastewater. A series of batch adsorption experiments was carried out in single and binary-metal solutions of Cd 2+ and Pb 2+ by changing the mixing proportions of SS and AAC grains. In addition, the effect of the liquid-to-solid ratio (L/S) on the removal of Cd 2+ and Pb 2+ in multi-metal solution was examined. Results showed that SS grains had a high afﬁnity with Cd 2+ in the single solution, while AAC grains had an afﬁnity with Pb 2+ . In the binary solution, the mixtures of SS and AAC grains removed both Cd 2+ and Pb 2+ well; especially, the tested adsorbents of SS+AAC [1:1] and SS+AAC [1:4] mixtures achieved approximately 100% removal of both metals. Based on the results in the multi-metal solutions, the metal removal % and selectivity sequence varied depending on the mixed proportions of SS and AAC grains and L/S values. It was found that the SS+AAC [1:1] mixture of SS and AAC grains showed 100% removals of Cd 2+ , Pb 2+ , Cu 2+ , Ni 2+ , and Zn 2+ simultaneously at L/S = 10 and 60.


Introduction
Due to rapid urbanization and industrialization, the discharge of wastewater containing toxic heavy metals from various sources such as industries, mines, vehicles, batteries, and metal-containing paints is increasing all over the world, especially in developing countries [1,2]. Heavy metal ions are non-biodegradable and tend to accumulate in living organisms, and those are considered toxic or carcinogenic ions [3,4]. Over the past five decades, for example, the annual global release of heavy metals reached 22,000 tons for cadmium, 939,000 tons for copper, 783,000 tons for lead, and 1,350,000 tons for zinc [5]. Inadequate treatments of toxic heavy metals such as Cd 2+ and Pb 2+ cause not only serious surface and groundwater pollutions and soil contamination [6][7][8] but also various human diseases including acute or chronic poisoning, dermatitis, brain damage in children, and digestive tract cancer [9][10][11]. Furthermore, the direct discharge of Cd 2+ and Pb 2+ into the sewage system causes a negative impact on the effectiveness of biological wastewater treatment [12]. Thus, Cd 2+ and Pb 2+ can be considered the most commonly available and most harmful heavy metals to humans as well as the environment. Thus, proper treatment of these heavy metals in wastewater is essential to protect public health and the environment.
Numerous efforts are being made to develop cost-effective innovative methods of wastewater treatment, such as hydroxide precipitation [13], membrane filtration [14], re-was carefully investigated in the multi-metal solutions (coexistence of Ni 2+ , Cu 2+ , and Zn 2+ ) at different L/S ratios ranging from 5 to 250.

Adsorbents Preparation and Characterization
Commercially available steel slag (SS) for civil engineering applications (Nippon Steel Cooperation and Sumitomo Metal Industries, Ltd., Saitama, Japan) and autoclaved aerated concrete (AAC) (Asahi Kasei Construction Material Corp., Tokyo, Japan) were used. General information on the manufacturing and characteristics of SS and AAC are given in Nippon Slag Association [66] and Trong et al. [67]. After crushing by hand in the laboratory, the tested samples were sieved to three grain fractions of <0.105, 0.105-2, and 2-4.75 mm. The dataset of material properties and the adsorption tests for AAC grains analyzed in this study were obtained from Kumara et al., 2019a [61].
The basic physical and chemical properties of SS and AAC grains are summarized in Table 1. The Brunauer-Emmett-Teller (BET) surface area (SSA) was measured by a ASAP2020 adsorption analyzer (Micromeritics, Norcross, GA, USA). Measured specific gravity (G s ) values of SS grains were higher compared to AAC grains, and pH values showed that SS grains are alkaline in water due to the hydration reaction of CaO with a release of OH-. The SSA values of SS grains were lower than those of AAC grains and decreased with decreasing grain size. This implied that SS grains had fewer inter-grain pores compared to AAC grains, and the outer-grain surface controlled the SSA of SS grains. The chemical composition of tested adsorbents was characterized by energy-dispersive X-ray spectroscopy (EDX; X-Max Extreme, Oxford Instruments, High Wycombe, UK) and X-ray diffractometry (XRD; XRD-7000, Shimadzu Cooperation, Kyoto, Japan). The EDX test data are shown in Table 2. Higher CaO (41.7%) and Fe 2 O 3 (22.5%), and lower SiO 2 (16.8%) were found for SS grains compared to those of AAC grains, indicating a potential for the ion exchange (Ca 2+ ) reaction with heavy metals in wastewater like other calcium silicate materials [68,69]. The XRD analysis showed that magnetite (Fe 3 O 4 ), iron oxidate (FeO), calcite (CaCO 3 ), calcium hydroxide (Ca(OH) 2 ), and halloysite (Al 2 Si 2 O 5 (OH) 4 ) were the main minerals in SS grains. Both EDX and XRD analyses confirmed the existence of high amounts of metal oxides and hydroxides in SS grains, implying that SS grains favored the adsorption of heavy metals such as Cd 2+ , Cu 2+ , and Pb 2+ [70][71][72].

Procedures of Batch Experiments
Stock solutions of Cd 2+ , Pb 2+ , Cu 2+ , Ni 2+ , and Zn 2+ (synthesized wastewater) were prepared by dissolving CdCl 2 , PbCl 2 , CuCl 2 , NiCl 2 , and ZnCl 2 , respectively, in deionized water (DI). The solutions' pH and ionic strength were controlled using 1N HCl, 1N NaOH, and NaNO 3 . All chemicals used for this study were analytical grade from Wako Pure Chemical Industries, Ltd., Osaka, Japan, with more than 98% chemical purity. A standard batch method recommended by the Organization of Economic Cooperation and Development [73] was used for all batch adsorption experiments. The concentrations of heavy metals and other ions were measured using a flame atomic absorption spectrophotometer (AAS; AA 6200, Shimadzu, Japan), the exact pH and EC values for each metal solution were measured before the experiment, and their changes were observed during the experiments using a pH/EC meter.
Test conditions of adsorption experiments are summarized in Table 3. The types of batch adsorption experiments in this study were categorized into four groups to investigate: (1) adsorption isotherm in a single-metal solution with L/S = 60 for Cd 2+ and 10 for Pb 2+ , (2) % removal in a binary-metal solution (Cd 2+ , Pb 2+ ) with L/S = 60, (3) % removal in multi-metal solution (Cd 2+ , Pb 2+ , Cu 2+ , Ni 2+ , Zn2+) with L/S = 60, and (4) the effect of L/S ratios on % removal in a multi-metals solution with different L/S ratios = 5-250. Three grain sizes of adsorbents of <0.105, 0.105-2, and 2-4.75 mm were used for the adsorption isotherm experiments in a single-metal solution, and an adsorbent grain size of 0.105-2 mm was used for other adsorption experiments. All test conditions were carried out with triplicate measurements, and the averaged values were given in the paper because of a small variation of measured data.

Adsorption Isotherms for Cd 2+ and Pb 2+ in a Single-Metal Solution
Adsorption isotherm experiments for Cd 2+ and Pb 2+ onto SS and AAC grains were carried out at natural pH with the initial metal concentration (C i ) of 0-5000 mg/L for Cd 2+ and 0-1500 mg/L for Pb 2+ (19 different concentrations, fully considering the actual metal concentration range in industrial wastewater [74]) at L/S = 60. To determine the maximum adsorption capacity and intensity of Cd 2+ and Pb 2+ onto the tested adsorbents, Langmuir (Equation (1)) and Freundlich (Equation (2)) models were used to fit the obtained experimental data: log Q e = log K f + 1/n (log C e ) where C e (mg/L) is the equilibrium concentration of heavy metals, Q e (mg/g) is the amount adsorbed per adsorbent at the equilibrium, b (g/L) is the Langmuir constant related to binding strength, Q m (mg/g) is the maximum adsorption capacity, K f (L/g) is the Freundlich adsorption capacity, and 1/n is the adsorption intensity.

Effect of Competitive Metal Ions on Cd 2+ and Pb 2+ Adsorption in Binary Metal Solution
To examine the effect of competitive metal ions on Cd 2+ and Pb 2+ adsorption, batch experiments were carried out in a binary-metal solution of Cd 2+ and Pb 2+ at natural pH with C i = 1000 mg/L. Five different mixtures of SS and AAC grains, i.e., mixing proportions of SS (alone), SS+AAC [4:1], SS+AAC [1:1], SS+AAC [1:4], and AAC (alone), were used as tested adsorbents. The ratios in brackets show the mixing proportions in weight % (e.g., (1:1) means a mixture of 50% SS and 50% AAC). In binary-metal adsorption experiments, the mixed solutions of Cd 2+ to Pb 2+ were used at metal molar ratios of 1:0, 1:0.25, 1:0.5, 1:0.75, 1:1, 1:2, and 1:5 to investigate the effect of Pb 2+ concentrations on Cd 2+ adsorption. Vice versa, the same mixing solutions of Pb 2+ and Cd 2+ solutions were used to investigate the effect of Cd 2+ on Pb 2+ adsorption. In each experiment, the metal removal percentage (R, %) was calculated using Equation (3) [46]:

Effect of Competitive Metal Ions on Cd 2+ and Pb 2+ Adsorption in Multi-Metal Solutions
To examine the effects of competitive metals on Cd 2+ and Pb 2+ adsorption onto tested adsorbents with different L/S, batch adsorption experiments in multi-metal solutions were carried out using a mixed metal solution of Cd 2+ , Pb 2+ , Cu 2+ , Ni 2+ , and Zn 2+ at natural pH with C i = 1000 mg/L. Like the adsorption experiments of binary-metal solutions, five different mixtures of SS and AAC grains (SS, SS+AAC [4:1], SS+AAC [1:1], SS+AAC [1:4], and AAC) were used with five different L/S values of 5, 10, 60, 100, and 250. In each experiment, the R value of each metal was calculated using Equation (3).

Adsorption Isotherms for Cd 2+ and Pb 2+ in Single Metal Solution
The measured adsorption isotherms for Cd 2+ and Pb 2+ onto SS grains with different grain sizes are shown in Figure 1. In the figures, the Langmuir model (Equation (1)) was well-fitted to the measured data except for the Cd 2+ adsorption onto SS grains with <0.105 mm (Figure 1a). The SS grain with <0.105 mm showed a very high Cd 2+ adsorption and did not show the maximum adsorption capacity (Q m ) in the range of C i = 0-5000 mg/L. For both Cd 2+ and Pb 2+ adsorptions, the adsorption decreased with increasing grain size. This can be understood to indicate that the adsorption surface area of tested adsorbents controlled the adsorption capacity, and the sample of <0.105 mm with higher SSA imposed a higher adsorption capacity compared to the samples with lower SSA of the grain sizes of 0.105-2, 2-4.75 mm ( Table 1). The Freundlich model (Equation (2)), on the other hand, fitted well with all tested adsorbents in the range of C i = 0-5000 mg/L for Cd 2+ and 0-1500 mg/L for Pb 2+ . This suggests that a monolayer adsorption is predominant at the early stage of the adsorption process (typically, C i < 1500 mg/L) and the adsorption process shifts to a multilayer adsorption, especially at higher than C i > 1500 mg/L, for both Cd 2+ and Pb 2+ [75,76].
The fitted adsorption parameters by Freundlich and Langmuir models for tested SS grains are summarized in Table 4 with the reported data for AAC grains [61]. It shows clearly that the SS grains have higher affinities (higher Q m and K f ) to Cd 2+ compared to AAC grains while the AAC gains have higher affinities to Pb 2+ compared to SS grains. These results imply that the mixing of SS and AAC grains would be effective for the simultaneous adsorption of Cd 2+ and Pb 2+ , as discussed in the following sections. For reference, the measured Q m values in this study were compared to the previously reported values for different types of adsorbents such as IBPs (including construction and demolition waste) and geo-and bio-sorbents and are summarized in Table 5. It is noticeably shown that SS has the highest Q m of Cd 2+ adsorption compared to those of other adsorbents.   Figure 2 shows the relationship between the amounts of Ca 2+ and adsorbed metal released onto SS grains sized 0.105-2 mm in the measured C i range. For both Cd 2+ and Pb 2+ adsorptions, the Ca 2+ was released linearly along with the metal adsorption (r 2 ≥ 0.88). The correlation regression showed that the ratios of released Ca 2+ became greater than 1 (2.24 for Cd 2+ and 6.37 for Pb 2+ ). This means that approximately 2 and 6 times of Ca 2+ were released when the one metal was adsorbed, indicating the adsorption mechanism was not controlled by a simple 1:1 ion exchange process of Ca 2+ and Cd 2+ /Pb 2+ on the adsorbent surface of SS grains in the single-metal solution system. On the other hand, Kumara et al. 2019a [61] observed an almost 1:1 relationship between the released Ca 2+ amount and adsorbed Cd 2+ /Pb 2+ for AAC grains, and the hydrated adsorbent surface was the dominant metal adsorption mechanism [77,78].

Removal of Cd 2+ and Pb 2+ in Binary Metal Solution
Measured values of the percentage of metal removed, R % (Equation (3)) for Cd 2+ and Pb 2+ in the binary-metal solution by tested adsorbents, were plotted against the Cd:Pb and Pb:Cd molar mixed ratios in Figure 3. Like the test results in a single-metal solution for the removal of Cd 2+ , SS had a strong affinity with Cd 2+ and became independent of the existence of Pb 2+ in binary-metal solution, SS, and its mixtures (SS, SS+AAC [4:1], SS+AAC [1:1], SS+AAC [1:4]) gave approximately R = 100% (Figure 3a-d,f-i). In contrast, the removal of Pb 2+ in the binary-metal solution was controlled by AAC, and AAC and its mixtures (AAC, SS+AAC [1:4], and SS+AAC [1:1]), which gave approximately R = 100% (Figure 3c-e,h-j). Previous studies reported that geo-and bio-adsorbents had difficulty achieving the sufficient removal of Cd 2+ in the presence of Pb 2+ in the binary-metal solution [30,31]. In this study, however, the tested results suggested strongly that the mixing of SS and AAC grains was effective to simultaneously remove Cd 2+ and Pb 2+ in a binarymetal solution; the mixtures of SS+AAC [1:1] and SS+AAC [1:4] were especially able to absorb Cd 2+ (or Pb 2+ ) completely, even in the presence of Pb 2+ (or Cd 2+ ) in the solution. In the batch adsorption tests in the binary-metal solution, the measured pH values after the adsorption (pH e ) ranged from 8-10 for AAC grains (Figure 3e,j) and 10-12 for SS and mixed adsorbents of SS and AAC (Figure 3a-d,f-i). This indicated that the metals were adsorbed onto the tested adsorbents under alkaline conditions. Previous studies suggested that the metal adsorption onto cementitious and adsorbents rich in calcium metal oxides/hydroxides resulted in combined chemical processes and reactions under alkaline conditions: i.e., hydration of the adsorbent surface, hydrolysis of metal ions, physisorption, chemisorption, ion exchange, and surface complexation and precipitation [33,61,77,[87][88][89][90]. For example, According to the metal speciation in pH-Eh diagram [91], Pb is presented as Pb 2+ or possible to precipitate as Pb(OH) 2 in the range of pH = 7-12 and most probably exists as Pb(OH) 3 in the range of pH >12. On the other hand, Cd exists mainly in the forms of Cd 2+ and Cd(OH) + in the range of pH = 9-13. As shown in Table 2, the SS and AAC grains tested in this study were rich in CaO and Fe 2 O 3 . Based on the results of previous studies and the chemical compositions of tested adsorbents, the simultaneous removal of Cd 2+ and Pb 2+ from the mixtures of SS and AAC grains in the binary-metal solution under alkaline conditions can be understood as shown in Figure 4. For SS grains, under the given experimental conditions, active hydroxides (-OH) on the surface of metal oxides/hydroxides (especially Fe 2 O 3 ) create a negative surface charge, resulting in the formation of an inner-sphere complex (surface complexation) by replacing Ca 2+ (high affinity with Cd 2+ ). For AAC grains, on the other hand, the ion exchange between Ca 2+ and Pb 2+ on the adsorbent surface becomes dominant, and the inner-sphere complex promotes the Pb 2+ adsorption onto the surface with a negative charge (high affinity with Pb 2+ ). Because these two reactions promote the release of Ca 2+ ions, the amount of Ca 2+ released became higher than the amount of metal adsorbed, as shown in Figure 2. In addition, hydrated CaO releases high amounts of OH-ions for both SS and AAC grains and promotes the precipitation of metal hydroxides (Cd(OH) 2 and Pb(OH) 2 ) under the alkaline condition.

Removal of Cd 2+ and Pb 2+ in Multi-Metal Solutions and Selectivity Sequence
Simultaneous removals of Cd 2+ and Pb 2+ in multi-metal solutions by the tested adsorbents were examined under different L/S ratios of 5, 10, 60, 100, and 250. The measured R values of metal ions in multi-metal solutions for SS, SS+AAC [1:1], and AAC are exemplified in Figure 5, and selectivity sequences are shown in Table 6. For all tested adsorbents, the R values of metals became high in low L/S conditions and decreased with increasing L/S. Especially, the R values became low at L/S = 60, 100, and 250, except for the Pb 2+ removal and the Cu 2+ removal of SS and AAC grains at L/S = 60 ( Figure 5). This suggests that control of L/S is an important factor to achieve high R values in multi-metal solutions, unlike the single-metal and binary-metal solutions.  As shown in Table 6, the selectivity sequences of metals in multi-metal solutions are highly dependent on the L/S for SS and AAC grains, and the Cd 2+ and Pb 2+ adsorptions are affected by the existence of competitive metal ions. The selectivity sequences for the mixtures of SS and AAC, on the other hand, were less dependent on the L/S compared to the results from single SS and AAC grains. For SS grains, the adsorption of Cd 2+ (and Cu 2+ , Ni 2+ , and Zn 2+ ) became much higher than that of Pb 2+ at low L/S (=5 and 10), and the Pb 2+ adsorption became higher than that of Cd 2+ . For AAC grains, the adsorption of Pb 2+ (and Cu 2+ ) became higher than that of Cd 2+ at all L/S conditions in accordance with the tested results from single-metal and binary-metal solutions. For the mixtures of SS and AAC (SS+AAC [1:1]), the adsorption of Cd 2+ became equal to the Pb 2+ adsorption at low L/S (=5 and 10), but the sequence became the opposite at high L/S (=60, 100, and 250). Based on the test results above, therefore, the mixtures of SS and AAC grains are able to simultaneously remove Cd 2+ and Pb 2+ in multi-metals solution with high R values of approximately 100% at low L/S conditions (5 and 10). Again, the test results suggest the control of L/S is a key factor for the practical application of the mixtures of SS and AAC grains as adsorbents to remove Cd 2+ and Pb 2+ in wastewater.
In this study, SS exhibited a high capacity to adsorb Cd 2+ from wastewater. For example, 0.105-2 mm particles of SS achieved effluent discharge standards of <0.001 mg/L for Cd 2+ up to Ci < 1500 mg/L by the single-batch adsorption under the experimental conditions. Additionally, the SS+AAC (1:1) mixture removed 100% of Cd 2+ and Pb 2+ from binary and multi-metal solutions. However, the following facts should be considered in future studies and applications for the sustainable use of these low-cost adsorbents. Along with the metal adsorption process, SS and AAC grains released a relatively high concentration of Ca 2+ . Additionally, SS grains showed high alkalinity in water compared to AAC. Therefore, those factors should be controlled before treated water is discharged into the natural environment.

Conclusions
The results of batch adsorption experiments of the tested adsorbents revealed that SS grains had a high affinity for Cd 2+ (>300 mg/g) but less affinity for Pb 2+ (<20 mg/g). In the binary-metal solution, the mixtures of SS and AAC grains, especially SS+AAC [1:1] and SS+AAC [1:4], removed 100% of Cd 2+ and Pb 2+ simultaneously without any effect of metal molar mixed ratios at the L/S ratio of 60. Remarkably, in the multi-metal solutions, the metal R and selectivity sequence varied depending on the mixing proportions of SS and AAC grains and L/S values. It was found that the SS+AAC [4:1] and SS+AAC [1:1] mixtures of SS and AAC grains removed 100% of Cd 2+ , Pb 2+ , Cu 2+ , Ni 2+ , and Zn 2+ simultaneously, at L/S = 10 and 60. The correlation regression showed that the ratios of released Ca 2+ became greater than 1, meaning that approximately 2 and 6 times of Ca 2+ were released in the onemetal adsorption process. For SS grains, active hydroxides (-OH) on the surface of metal oxides/hydroxides created a negative surface charge, resulting in the formation of an innersphere complex by replacing Ca 2+ . For AAC grains, the ion exchange between Ca 2+ and Pb 2+ on the adsorbent surface becomes dominant and the inner-sphere complex promotes the Pb 2+ adsorption onto the surface with a negative charge. Thus, the amount of Ca 2+ released became higher than the amount of metal adsorbed. Further studies are needed to examine the effects of other factors such as initial pH, temperature, and other competitive ions that control the heavy metal adsorption of SS and AAC grains. Additionally, the regeneration of metal-adsorbed adsorbents (i.e., collection of adsorption heavy metals from adsorbents) should be examined for practical application. The application of IBPs for the wastewater treatment, however, is essential with a high potential from the viewpoints of material efficiency and saving natural resource consumption.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.

AAC
autoclaved aerated concrete b Langmuir constant (g/L) BET Brunauer-Emmett-Teller CDW construction and demolition waste C e equilibrium concentration (mg/L) C i initial concentration (mg/L) DI deionized water EC electrical conductivity (mS/cm) EDX energy-dispersive X-ray spectroscopy G s specific gravity (-) IBPs industrial by-products K f Freundlich adsorption capacity (L/g) LOI loss on ignition (%) L/S liquid-to-solid ratio n adsorption intensity (-) pH e equilibrium pH Q e equilibrium adsorption (mg/g) Q m maximum adsorption capacity (mg/g) R removal percentage (%) SS steel slag SSA surface area (m 2 /g) w AD gravimetric water content at air-dry [g/g, in %] XRD X-ray diffractometry