Characterization of the adsorbent
Granulation analysis was performed and the results are as follows: 0–0.212 mm − 13.1%; 0.212–0.5 mm − 73.8%; 0.500–1.0 mm − 11.2%; 1.0–1.7 mm − 0.51%; > 1.7 mm − 1.39%. Based on these studies, it was found that the CFBC-S particles are heterogeneous and their diameter influences the course of adsorption. According to literature reports, smaller slag particles are characterized by a larger specific surface area and the number of active centers, which translates into higher efficiency of sorption processes of metal ions (Kostura et al. 2017). These data influenced the decision to use the smallest fractions with a diameter less than 0.212 mm.
The analysis of particle size distribution showed one peak at the particle size of 955.4 nm (Fig. SM1, supplementary material). The analysis was difficult to carry out because some of slag particles did not suspend in the aqueous solution, and larger ones fell to the bottom of the suspension. Hence, it was only possible to analyze suspended samples.
Bulk density measurements were carried out by loosely filling slag into a cylinder and by compaction on a vibrating table. The results were estimated at 0.82 and 1.34 g/cm3, respectively. The results of the analysis may be useful from the point of view of industrial application in building and construction materials.
Elemental analysis using the SEM-EDS method was performed and the results are shown in Ta ble 1 and Fig. SM2 (supplementary material). As it is seen, slag mainly includes following elements: Ca, O, P, Si, Al, Mg, Fe, C. The method is based on a spot measurement on the sample surface. Slag is a complex mixture, hence the quantitative and qualitative composition may slightly differ in different places of the agglomerates due to the location of the measuring point.
Table 1
Elemental composition of CFBC-S (SEM-EDS analysis)
Elements
|
C
|
O
|
Na
|
Mg
|
Al
|
Si
|
P
|
S
|
K
|
Ca
|
Ti
|
Mn
|
Fe
|
CFBC-S [%], weight
|
1.33
|
43.8
|
0.74
|
2.51
|
3.54
|
3.74
|
14.2
|
0.46
|
0.19
|
27.1
|
0.44
|
―
|
1.89
|
CFBC-S [%], atomic
|
2.5
|
61.6
|
0.72
|
2.33
|
2.95
|
2.99
|
10.3
|
0.32
|
0.11
|
15.2
|
0.21
|
―
|
0.76
|
Oxides
|
CO2
|
Na2O
|
MgO
|
Al2O3
|
SiO2
|
P2O5
|
SO3
|
K2O
|
CaO
|
TiO2
|
MnO
|
Fe2O3
|
CFBC-S [%]
|
4.89
|
1.0
|
4.17
|
6.68
|
7.99
|
32.48
|
1.14
|
0.23
|
37.98
|
0.73
|
-
|
2.71
|
The BET and BJH analysis revealed the following results: specific surface area 1.87 m2/g, pore volume (Vp) 0.0096 cm3/g and mean pore diameter (Apd) 21.2 nm. The shape of the obtained adsorption isotherms resembles the type III isotherm, which is related to cooperative adsorption (Figs SM3 – SM6, supplementary material). This means that previously adsorbed particles can lead to increased sorption of the remaining particles. The interaction of copper ions with the slag adsorbent has a positive effect on the adsorption of remaining copper ions when the ions have already been adsorbed at least once. The result is a bulging isotherm towards the pressure axis.
SEM images of CFBC-S particles are presented in Fig. 1. The slag particles are irregular in shape, compact, spongy and have a porous surface. The irregular shape depends on temperature and duration of a combustion process. More crystalline and spherical particles result from a longer process (Liu et al. 2014). Figure 2 presents TEM images of the samples showing flocs of irregular shape and different sizes. The particles are of different shades, the darker zones correspond to the thicker material. Similar observations were found in the literature (Li et al. 2011; Assi et al. 2020).
FT-IR studies of CFBC-S before and after Cu(II) ions adsorption were performed and the spectra are shown in Figure SM7 (supplementary material). To this purpose, sorption frequencies of functional groups, functional bonds and types of vibrations were identified. As it is seen, there is an increase in the intensity of peaks after Cu(II) adsorption observed. The following peaks have been observed: 593, 416, 406, 390 cm− 1 (bending vibrations Si-O-Si), 677, 611 cm− 1 (stretching vibrations Al–O), 713 cm− 1 (symmetric stretching of Si–O–Si and Al–O–Si). 874 cm− 1 (symmetric stretching of Al–O–M, vibration of carbonates (calcite)), 113 cm− 1 (asymmetric stretching vibrations of silica Si–O–Si and Al–O–Si), 1408 cm− 1 (valence vibration of carbonate ions), 3252 cm− 1 (stretching vibrations O–H), 3643 cm− 1 (asymmetric and symmetric stretching vibrations O–H (probably amorphous silicates or hydrated aluminum silicates)). The wide band at around 3600 − 3200 cm− 1 appeared after, but the weak sharp peak at 3643 cm− 1 disappeared. The peaks at 594, 611, 677, 713, 874, 1113, 1408 cm−1 became more intensive due to a probable complexation reaction and formation of surface complexes with Cu(II) ions or bonds with the metal ions (Cu–O) (Kavaliauskas et al. 2015; Ueda et al. 2000; Iliashevsky et al. 2016).
Adsorption studies
Effect of initial pH
The impact of initial pH on the adsorption efficiency was analyzed and the results are shown in Fig. 3 and (Fig. SM8, supplementary material). The following conditions were applied in the experimental procedure: initial pH range of 1.8–5.6, initial concentration of Cu(II) ions 100 mg/L, adsorbent dosage 1–5 g/L, contact time 60 min, temperature 23°C, agitation speed 200 rpm. As it is seen, high adsorption efficiency was reported at initial pH 1.8 for the slag doses of 1 g/L (86.6%), 3 g/L (90.7%), 5 g/L (98.8%). An increase in initial pH resulted in a gradual decrease in adsorption efficiency. The experimental adsorption capacity also reported a stable decrease in the range of 19.5–14.7 mg/g. In considering the influence of the initial pH, the surface charge of the slag and the degree of speciation should be taken into account. The tested adsorbent is alkaline in nature, therefore it may increase the pH in an aqueous solution during adsorption. The presence of such anions in the adsorbing material as SiO32−, CO32−, PO43−, OH−, SO42− may contribute to the precipitation of copper ions from the solution at higher pH. The pH parameter influences electrostatic charge of the metal oxides present in the adsorbent material. Hence, it is highly probable that the adsorption of Cu(II) ions may be associated with ion exchange and/or complexation by bonding with oxygen groups. The probable ion exchange mechanism between H+ and Cu2+ ions can be proposed by the equations 1–3.
$$X\text{}OH+ {H}_{3}{O}^{+}\leftrightarrow X\text{}O{H}_{2}^{+}+ {H}_{2}O$$
1
$$X\text{}OH+ {OH}^{-}\leftrightarrow X\text{}{O}^{-}+ {H}_{2}O$$
2
$$2\left(X\text{}{O}^{-}\right)+ {Cu}^{2+}\leftrightarrow {\left(X\text{}O\right)}_{2}Cu$$
3
where: X may be Fe, Si, Al or another element. It should be pointed out that the proposed mechanism was not confirmed by additional experiments in this research (Kalak and Cierpiszewski 2019).
Effect of sorbent dosage
The impact of CFBC-S dosage on the adsorption of Cu(II) has been studied and the results are presented in Fig. 4. The experiments were conducted under following conditions: initial pH 1.9, initial concentration 100 mg/L, contact time 60 min, temperature 23°C, agitation speed 200 rpm. The results showed a rapid increase in adsorption efficiency up to 98% with the use of slag dosage 0.25–5 g/L. The dose of 5 g/L is considered optimal under these experimental conditions. The addition of higher doses of CFBC-S did not cause any significant changes and the process efficiency remained at the same level. Moreover, analysis of adsorption capacity revealed that: firstly, it increased up to 2.58 mg/g (dosage 3 g/L) and secondly, it gradually decreased to 0.4 mg/g at a dose of 25 g/L (Fig. SM9, supplementary material). At higher slag doses, the active centers available for further adsorption were not fully utilized, hence a decrease in sorption capacity was observed (Liu et al. 2014). Consequently, the optimal experimental adsorption capacity can be estimated in the range between 2.0 and 2.6 mg/g.
Effect of initial concentration of Cu(II) ions
The influence of the initial concentration of Cu(II) ions was investigated and the results are shown in Fig. 5. Based on the previous research results, the following experimental conditions were used: initial concentration of Cu(II) (2.5–100 mg/L), adsorbent dosage 2–5 g/L, initial pH 1.9, contact time 60 min, agitation speed 200 rpm, T = 23°C. The adsorption curves show an upward trend in all cases of slag doses. Higher adsorption efficiencies were obtained at the initial concentration of 100 mg/L (98.2% − 5 g/L of slag, 91.8% − 4 g/L, 89.0% − 3 g/L, 87.9% − 2 g/L). An increase in experimental adsorption capacity was also observed (Fig. SM10, supplementary material). This is due to the fact that free active sites were still available, and their total available amount influenced the adsorption efficiency.
Effect of contact time
The research results on the influence of contact time on adsorption efficiency and adsorption capacity are shown in Figs. 6 and 11 and Figure SM11 (supplementary material). The study of the contact time in adsorption is important from the point of view of potential industrial use. Determining optimal contact time can help to design processes efficiently and reduce costs. The following experimental conditions were used in these studies: initial concentration of Cu(II) ions 100 mg/L, initial pH 1.9, slag dosage 1–5 g/L, temperature 23°C, agitation speed 200 rpm. As shown in Fig. 6, the maximum level of adsorption efficiency was achieved in the range of 20–30 minutes of the process and no changes were observed until 60 minutes, hence there was no need to continue the experiment for a longer time. An increase in adsorption efficiency in the first stage may be a consequence of the availability of a large number of free active sites on slag surface and a high concentration of Cu(II) ions at the interface. The gradual occupation of active centers by metal ions during mixing contributed to equilibrium in the system.
Kinetic models
The kinetics of Cu(II) adsorption on CFBC-S was analyzed with pseudo-first order (PFO) and pseudo-second order (PSO) models. The calculation results (reaction rate constant k, equilibrium adsorption capacity qe and correlation coefficients R2) are presented in Table 2, and the plots are shown in Figures SM12 and SM13 (supplementary material). The performed calculations showed that higher coefficients R2 were recorded in the case of PSO model, which is related to greater correlation between the experimental qe and the calculated qt values. Therefore, it can be concluded that the kinetics of Cu(II) adsorption on CFBC-S is better described by the PSO model. Chemisorption and electrostatic attraction on the adsorbent surface could take place during these processes.
Table 2
Kinetic parameters of pseudo-first order and pseudo-second order models
Adsorbent
|
Adsorbent dosage [g/L]
|
PFO kinetic model
|
PSO kinetic model
|
kad
[min− 1]
|
qe
[mg/g]
|
R2
|
k
[g/mg min]
|
qe
[mg/g]
|
R2
|
CFBC-S
|
3
|
0.134
|
9.449
|
0.971
|
0.008
|
24.110
|
0.999
|
4
|
0.130
|
10.223
|
0.934
|
0.010
|
22.332
|
0.999
|
5
|
0.118
|
7.229
|
0.958
|
0.012
|
19.711
|
0.999
|
Isotherm models
Langmuir and Freundlich isotherms were used to analyze the studied adsorption process. The calculations of isotherm parameters are shown in Table 3 and the isotherms are included in Figures SM14 and SM15 (supplementary material). Based on the Langmuir equation, the maximum adsorption capacities were calculated and are as follows: 53.04 mg/g (CFBC-S dosage 2 g/L), 56.05 mg/g (3 g/L), 60.92 mg/g (4 g/L) and 70.34 mg/g (5 g/L). According to the calculated correlation coefficients R2, adsorption reactions carried out in these studies are more closer to the Freundlich model than to the Langmuir model.
Table 3
Parameters of Langmuir and Freundlich isotherm models
Adsorbent
|
Adsorbent dosage [g/L]
|
Langmuir isotherm
|
Freundlich isotherm
|
Calculated qm [mg/g]
|
KL [L/mg]
|
R2
|
Kf
[mg/g] [L/mg](1/n)
|
n
|
R2
|
CFBC-S
|
2
|
53.04
|
0.049
|
0.824
|
1.315
|
0.986
|
0.933
|
3
|
56.05
|
0.037
|
0.888
|
1.213
|
0.938
|
0.938
|
4
|
60.92
|
0.031
|
0.910
|
1.171
|
0.906
|
0.940
|
5
|
70.34
|
0.019
|
0.929
|
1.049
|
0.964
|
0.950
|