Introduction

Water is an essential resource for man and his environment; for this reason, its contamination has become one of the challenges that the world must meet for its well-being [1,2,3]. Indeed, the presence of strong industrialization, the multiplication of urbanization, and non-compliance with regulations regarding discharges has had an alarming impact on water quality [4]. Heavy metals, dyes, organic matter, and biocides are among the pollutants of major concern [5]. Heavy metal ions are among the most released contaminants, and for this reason, they are particularly worrisome [6]. Heavy metals are not biodegradable; therefore, they tend to bioaccumulate, which is their overtime increase of concentration in living organisms [1, 2, 7].

Hence, it is unavoidable to remove the metal ions from water environment by efficient and rigorous methods such as chemical precipitation [8], coagulation/flotation [9], membrane technologies [10], ion exchange [11], electrochemical technologies [12], and adsorption phenomenon [13,14,15,16]. Among these techniques, adsorption is still considered as one of the better technique used to eliminate metal ions from aqueous medium because of its simple operation, high efficiency, and low cost [17, 18]. In this work, we discuss the possibility of eliminating the nickel contained in water by adsorption phenomenon using the slag from the blast furnace (BF) of the El-Hadjar Annaba/Algeria complex as an adsorbent. BF slag is a solid waste material which is by product of steel making industry; this co-product is formed during the development of cast iron from iron ores. In the world, the annual BF slag production produced is in the tens of million tons, which represents a real environmental problem [19, 20]. From to the literature, many researchers have demonstrated the effectiveness of blast furnace slag in the area of adsorption in solution [21, 22]. Nickel is considered among the most toxic elements of the heavy metals, as it is not biodegradable and thus considered as a risk factor for human health [20, 21]. The main sources of nickel are the industrial discharges from diffuse agricultural, metallurgy, oil refining, exploitation of mineral deposits, electroplating, paint formulation, and the battery manufacturer. In this work, the analyses by X-ray fluorescence, the analyses by X-ray diffraction, and the measurement of the specific surface area by BET method of slag were realized. Furthermore, the contact time, stirring speed, pH, temperature, particle size, and initial concentration were done to evaluate the removal efficiency of the nickel by BF slag as an adsorbent. Adsorption isotherms were presented by Langmuir and Freundlich models. Adsorption kinetic was utilized to describe the kinetics and mechanisms of adsorption process. The thermodynamic parameters were investigated to study the energy exchanges. The purpose of this work is to present a reliable adsorbent at reasonable prices to contribute at the birth of a venerable and sustainable environment.

Methods/experiment

Treatment of BF slag

The samples of the BF slag were collected at the El-Hadjar steel complex Annaba/Algeria, in the form of rocks and dark color. The washing performed at the crude slag is represented as follows:

  • The considered BF slag samples were washed with distilled water and air dried for 48 h

  • From an automatic grinder, the washed samples were grinded according to our request: 200, 300, 400, and 500 μm

  • The crushed samples were sieved with different the particle sizes (200, 300, 400, and 500 μm)

  • Samples with different grain sizes were separated, washed with distilled water, steamed at 105 °C, and stored in plastic boxes

Analytic methods

The nickel concentration was measured by atomic absorption spectroscopy method (PerkinElmer 3110) equipment. The pH of solution was adjusted with a digital pH meter (Hanna Instruments). The characterization was carried out by X-ray fluorescence (Siemens SRS 3000) and X-ray diffraction (Rigaku Ultim IV). The grinding of the slag was carried out by a planetary crusher (Fritsch pulverisette 7 premium line).

The specific surface areas of the BF slag particles were obtained by using the nitrogen gas adsorption-desorption method. The isotherm data for nitrogen gas desorption at 77K were analyzed with the Brunauer, Emmett et Teller model (model BET).

Adsorption protocol

A series of batch experiments were conducted to study the kinetics, mechanism, isotherms, and thermodynamic parameters in the nickel adsorption in solution by BF slag. The study of adsorption kinetics was carried out by adding 1 g of BF slag in aqueous solutions prepared with low nickel nitrate (Ni(NO3)2, 6H2O) in 1 l capacity beakers. The continuous mixing of the solution was ensured during all the tests by mechanical stirrer at different speeds. The temperature was controlled with a water bath equipped with a thermostat. The pH solution was fixed in desired value adding a few drops of concentrated ammonia. The adsorption kinetics were followed by a 5-ml sample using a graduated pipette fitted with a paper filter every 10 min. The samples taken were stocked in flasks, and the concentration of Ni (II) ions was measured using atomic absorption spectrometry (AAS). The experimental protocol was carried out on 04 identical workstations. Two samples were taken from each station to maintain the same concentration. The 04 workstations started at the same time in order to ensure continuity and respect the agitation time. After each two samples, the solution containing nickel and slag is replaced with another identical one. It is obvious that the agitation time is respected and goes in the continuity.

The amounts of adsorbed Ni (II) ions at equilibrium (qe) and at time t (qt) were calculated from the mass balance expression given by:

$$ {q}_e=\frac{C_{0-}{C}_e}{m}\times V $$
(1)
$$ {q}_t=\frac{C_{0-}{C}_t}{m}\times V $$
(2)

where C0 is Initial solute concentration (mg/L), Ce is the residual solute concentration at equilibrium (mg/L), Ct is the residual solute concentration at time t (mg/L),V is the volume of the solution (L), and m is the mass of adsorbent (g).

Results and discussion

Characterization of BF slag

The chemical composition of BF slag is reported in Table 1, where we find a dominance of the SiO2 (41.1%), CaO (37.2%), and Al2O3 (8.2%). The remaining oxides represent relatively low levels (Fe2O3, NiO, MgO, K2O, Na2O).

Table 1 Chemical composition of BF slag
Full size table

Figure 1 represents the diffractogram of blast furnace slag sample from the El-Hadjar complex. The investigations carried out showed that the slag principally lime contains silica and at lesser degree of alumina. This observation justified the results obtained by X-ray fluorescence. The result of the investigations has shown that the specific surface is 275.8 m2.g−1.

Fig. 1
figure 1

Diffractogram of BF slag sample from the El-Hadjar complex

Full size image

Adsorption study

Determination of adsorption equilibrium time

To determine the pseudo-equilibrium of the nickel adsorption of the blast furnace slag, we varied the time interval from 0 to 300 min. Figure 2 showed the impact of the contact time on the adsorption capacity of nickel in solution beneath our experimental conditions: C0 30 mg/L, Vag. 200 rpm, pH 5.3, T 20 °C, qeexp 400 μm, and Msolid 1 g.

Fig. 2.
figure 2

Residual concentration of Ni (II) as function of time

Full size image

According to kinetic study of the Ni(II) adsorption, we have noticed that the shape of the curve is subdivided into three phases: fast phase, slow phase, and permanent phase (Fig. 2).

For the first phase which varies from 0 to 50 min, the nickel elimination efficacy is very elevated; this is due to the numbers of adsorption vacant sites and probably to the specific surface of the blast furnace slag which could be negatively charged (adsorption under the effect of electrostatic attraction) [23,24,25,26]. On the other hand, for the second phase (50 to 90 min) the adsorption efficiency gradually decreases since the vacant sites are not as available.

During the last phase, the adsorption capacity has become constant. This means that the adsorbent has reached the saturation phase. For this purpose, we consider 90 min as the equilibrium contact time [27, 28].

Effect of initial pH

The pH of the solution is an important parameter in the adsorption phenomenon because it affects the form of metal in solution, as well as the surface properties of the adsorbent [29, 30]. In this work, three different media were used: pH 2, pH 4.5, and pH 5.3; C0 30 mg/L; Vag. 200 rpm; T 20 °C; qeexp 400 μm; and Msolid 1 g. (Fig. 3).

Fig. 3
figure 3

Effect of initial pH

Full size image

The kinetic study showed at pH 2 the adsorption is unfavorable (Figs. 3 and 4). At low initial pH value, competition is observed between H+ ions and Ni++ ions in the solution a given that the ionic radius of hydrogen ions is much smaller than the ionic radius of nickel ions. Therefore, it is for this reason that hydrogen ions are the most adsorbed [31, 32].

Fig. 4
figure 4

Nickel removal efficiency as a function of pH

Full size image

As the initial pH increased from pH 2 to pH 4.5, the adsorption capacity increased from 12.91 to 23.48 mg/g, and the yield increased from 43.03 to 78.26% (Figs. 3 and 4). Indeed, the efficiency of adsorption can be explained by the effect that the pH of the solution has favorably influenced the surface charges of adsorbent [33]. Moreover, when the initial pH equals 5.3, the nickel ions precipitated owing to augmentation in hydroxide ions in aqueous solution which decrease the adsorption yield [34]. Accordingly, pH = 4.5 value was selected as optimum pH for Ni(II) adsorption by BF slag in aqueous medium.

Effect of temperature

In adsorption processes, the temperature of the medium is a very influential parameter because it contributes to the fixation of metal ions on the surface of the adsorbent [35, 36]. The optimum temperature was set by realizing out adsorption tests at different temperatures (20.40 and 50 °C) under the following experimental conditions (C0. 30 mg/L; Vag 200 rpm; pH 4.5; qeexp 400 μm; Msolid 1 g), initial pH 4.9, agitation speed 200 rpm, adsorbent mass 1 g, initial concentration 30 mg/L, solid particle size 400 μm, and contact time 90 min (Fig. 5). The study of the temperature influence on the nickel adsorption has shown an inverse relationship between temperature and quantity of adsorbed ion, and it revealed that the adsorption is best at 20 °C (Figs. 5 and 6). Indeed, the nickel removal yield by BF slag between 20 and 40 °C, 20–50 °C, and 40–50 °C decreased respectively 16.8%, 21.96%, and 5.16% (Fig. 6).

Fig. 5
figure 5

Effect of temperature

Full size image
Fig. 6
figure 6

Nickel removal efficiency as a function of temperature

Full size image

From this study part, we deduced that the adsorption of nickel in aqueous solution by BF slag is exothermic [37]. The rise of the temperature of the medium has increased the resistance to mass transfer, and consequently, it slowed the diffusion of metal ions towards the surface of the adsorbent. This resistance could be due to the exothermic nature of adsorption processes of the nickel by BF slag [38].

Effect of particle size

The adsorbent granulometry has an important role in the rate of transfer of metal ions from the solution to the adsorbent [39, 40]. In this work, we have varied the particle size of BF slag from 200 to 500 μm by following the same experimental protocol (C0. 30 mg/L; Vag 200 rpm; pH 4.5; T = 20 °C; M 1 g) (Fig. 7).

Fig. 7
figure 7

Effect of particle size

Full size image

According to the experimental results conducted, it was noticed that when granulometry of the solid decreased from 500 to 300 μm, the amount adsorbed and yield respectively increased from 19.72 to 27.68 mg/g and 65.73 to 92.26% (Figs. 7 and 8). This means that the adsorption loses its effectiveness with the increase of the granulometry. This phenomenon is probably due to the narrowing of the specific surface of the adsorbent [41, 42]. It is important to clarify that with granulometry equal to 200 μm, we observed the decrease in the adsorption capacity and removal efficiency of 12.15 mg/g and 40.5% (Figs. 7 and 8). This result could probably be clarified by the emergence of the phenomenon of coalescence, that is to say the gathering of the particles of the adsorbent and the return to the upper diameters [42] (Figs. 7 and 8).

Fig. 8
figure 8

Nickel removal efficiency as a function of particle size

Full size image

Effect of initial concentration

The nickel adsorption by BF slag in optimum conditions for different synthetic solutions has been studied. In this step, we noticed the evolution in the adsorbed quantity from 9.26 to 53.58 mg/g by varying the initial concentration from 10 to 100 mg/L (Fig. 9). Indeed, as the initial concentration increases, the probability of contact between the adsorbent and the adsorbate per unit volume increased; therefore, the adsorbed amount increased [43, 44]. Nevertheless, the number of surface adsorption sites of the adsorbent material is limited; so, the adsorption capacity becomes constant [45,46,47]. Indeed, the plateau values showed that higher concentrations (> 80 mg/L) had no effect on the adsorption capacity and thus in the adsorption process.

Fig. 9
figure 9

Effect of initial concentration on nickel adsorption

Full size image

The value of the adsorbed quantity at tier level is 53.58 mg/g. This result allows us to conclude that this quantity represents the maximum adsorption capacity, which can be fixed by 1 g of BF slag under our experimental conditions. It is important to indicate that at lower concentrations, the adsorption efficiency is higher due to the availability of vacant sites (Fig. 10).

Fig. 10
figure 10

Nickel removal efficiency as a function of initial concentration

Full size image

Adsorption isotherm

The adsorption isotherm indicates the interaction between adsorbate and adsorbent, and it is brief to find the best fit in the adsorption process. In this study, the nickel adsorption by BF slag from the solution was fitted to both the Langmuir and Freundlich isotherm.

The Freundlich model is well adapted to describe the equilibrium in aqueous phase and describe that the multilayer adsorption is processed on a heterogeneous surface [48]. Its empirical formula is:

$$ {q}_e=\frac{x}{m}=k.{C}_e^{1/n} $$
(3)

where qe is the amount of adsorbate fixed at equilibrium by the adsorbent (mg g−1), Ce is the residual concentration at equilibrium (mg L−1), and KF and 1/n are the Freundlich constants related to adsorption and affinity. The linearized Freundlich relation is written as follows:

$$ \mathit{\ln}{q}_e= logk+\frac{1}{n}\mathit{\ln}{C}_e $$
(4)

The model of Langmuir [49] is most frequently employed to present the data on adsorption from solution. It is represented by the following equation:

$$ {q}_e=\frac{b.{q}_m.{C}_e}{1+b.{C}_e} $$
(5)

where qe is the amount of adsorbent fixed at equilibrium by the adsorbent (mg g−1), Ce is the residual concentration at equilibrium (mg L−1), qmax is the maximum capacity (mg g−1), and b is the thermodynamic constant of the adsorption equilibrium (L.mg−1). The linear form of the Langmuir equation is shown as follows (Eq. 6):

$$ \frac{C_e}{q_e}=\frac{1}{q_{max}}{C}_e+\frac{1}{q_{max}b} $$
(6)

The theoretical parameters of adsorption isotherms along with regression coefficients are listed in Table 2. Figures 11, 12, and 13 show the Langmuir model, Freundlich model, and adsorption isotherm of the nickel.

Table 2 Isotherm parameters nickel adsorption by BF slag
Full size table
Fig. 11
figure 11

Presentation of Freundlich model

Full size image
Fig. 12
figure 12

Presentation of Langmuir model

Full size image
Fig. 13
figure 13

Adsorption isotherm presentation

Full size image

According to Fig. 13, the adsorption isotherms presented a classic appearance of a type I isotherm. The adsorption takes place gradually until it reaches a saturation level. According to the Table 2, the correlation coefficient value (R2) of the Langmuir model was higher than that of the Freundlich model (R2 Langmuir = 0.99, R2 Freundlich = 0.87), and the value of nickel adsorption capacity obtained from the Langmuir model was close to the experimental value. This implicate that the Langmuir equation was more appropriate to represent the process of nickel adsorption by BF slag in aqueous solution [23, 50].

According to Table 3, we also noticed that the value of the heterogeneity factor is between 1 and 10 (n = 2.77) which showed that the nickel absorption is favorable [51, 52].

Table 3 Kinetics data the nickel adsorption of by BF slag
Full size table

The dimensionless RL parameter is a feature of the Langmuir isotherm model that can illustrate the type of adsorption process [53, 54]. This parameter RL is determined by the following equation:

$$ {R}_L=\frac{1}{1+{C}_0b} $$
(7)

where b is the Langmuir isotherm constant, and C0 is the initial concentration of solution.

The RL values indicate the nature of the adsorption according to the following assumptions: favorable (0 < RL < 1), unfavorable (RL > 1), linear (RL = 1), and irreversible (RL = 0) [55, 56]. According to Fig. 14, we noticed that the values of the parameter RL are between 0.237 and 0.03, which implies that the nickel adsorption process is favorable and follows a monolayer adsorption [56, 57]. According to Fig. 14, we noticed that the values of the parameter RL are between 0.237 and 0.03, which implies that the nickel adsorption process is favorable and follows a monolayer adsorption [56, 57].

Fig. 14
figure 14

Evolution of RL ration as a function of initial concentration

Full size image

Adsorption kinetics

The adsorption kinetics of nickel by the BF slag was investigated using pseudo-first-order and pseudo-second-order kinetics at different temperatures under optimal experimental conditions (Figs. 15 and 16). The pseudo-first-order model is represented by Lagergren equation [58]:

$$ \mathit{\log}\left({q}_e-q\right)=-{k}_{Lag}t+\mathit{\log}{q}_e $$
(8)
Fig. 15
figure 15

Pseudo-first-order kinetic

Full size image
Fig. 16
figure 16

Pseudo-second-order kinetic

Full size image

where qe is the adsorbed quantity at equilibrium (mg/g), q is adsorbed quantity at time t (mg/g), and t is the time of adsorption process; in this study, it is from 0 to 90 min; kLag is the constant pseudo-first-order sorption speed (s−1). The pseudo-second-order is represented by the Blanchard equation [59]:

$$ \frac{t}{q}=\frac{1}{k_b{q}_e^2}+\frac{1}{q_e} $$
(9)

where qe is the adsorbed quantity at equilibrium (mg/g), q is the adsorbed quantity at time t (mg/g), and t is the time of adsorption process; in this study, it is from 0 to 90 min; kb is the constant of pseudo-second-order sorption speed (min−1).

The curves obtained by fitting the pseudo-first-order and pseudo-second-order kinetics are presented in Fig. 15 and Fig. 16, respectively, and the relevant parameters are represented in Table 3.

According to Table 3, the values of the correlation coefficients R2 for the temperature studied were very high (R2 ≥ 0.98) for the Blanchard model; on the other hand, the values of the correlation coefficients of Lagergren model were less important R2 ≥ 0.96.

Also, the maximum values of the calculated adsorption capacity determined using the kinetic models were compared with experimental values. The results indicated the maximum adsorption capacity determined using the Blanchard model was closer to the experimental values. These results showed that the adsorption of nickel follows the kinetics of the pseudo-second-order model.

Adsorption mechanism

Weber and Morris [60] reported that if intra-particular scattering is involved in the sorption process, by increasing the adsorbed amount as a function of the square root of time, we need to obtain a line. This step is limiting if the line passes through the origin [61]. In the case where these lines do not pass through the origin, this indicates that the diffusion in the pores is not the only limiting mechanism of the sorption kinetics. It appears that other mechanisms are involved [62, 63]. The relation of Weber and Morris is presented as follows:

$$ q={k}_w\sqrt{t}+C $$
(10)

where q is the quantity adsorbed at time t, t is the time measured in minute, kw is the diffusion rate constant in the pores (mg/m. min ½), and C is the intercept .

Kw and C are obtained from the slope of the straight line of q versus t1/2 (Fig. 17). The internal transport parameters values are represented from Table 4.

Fig. 17
figure 17

Intra-particle diffusion kinetic

Full size image
Table 4 Values of transport parameters
Full size table

Optimization of kinetic data with the use of Weber and marries model confirmed that internal diffusion is not the main factor determining the process of nickel adsorption on BF slag (Fig. 17). Indeed, we noticed that the value of intercept is nonzero (C ≠ 0). It is important to note that the larger the C value, the greater the influence of the boundary layer on adsorption, and the greater the dependence of the adsorption [64, 65]. This result allowed us to conclude that the adsorption process is controlled as the first step by the external transport, which is an instantaneous process followed by intra-particle diffusion [66, 67].

Thermodynamic study

The variations in standard free energy (ΔG), standard enthalpy (ΔH), and standard entropy (ΔS) were used to speculate on the adsorption mechanism. These thermodynamic parameters are calculated by utilizing the Equations 11, 12, and 13 [68,69,70]:

$$ \Delta {G}^0=- RTln{k}_d $$
(11)
$$ \mathit{\ln}{k}_d=\frac{\Delta {H}^0}{R}\times \frac{1}{T}+\frac{\Delta {S}^0}{R} $$
(12)

The distribution coefficient is calculated from Equation 11 [71, 72].

$$ {k}_d=\frac{C_i-{C}_e}{C_e}\times \frac{V}{M}=\frac{q_e}{C_e} $$
(13)

Figure 18 shows the ln kd versus 1/T plot for the adsorption of the nickel at different temperatures. The values of the thermodynamic parameters are grouped in Table 5. The negative value of Gibbs energy (ΔG) showed that the adsorption process is spontaneous and realizable [73, 74]. The negative value of the enthalpy variation (ΔH) indicates that the adsorption is exothermic and is physical in nature involving weak forces of attraction at solid/liquid interface [37, 75]. The value negative of entropy (ΔS) indicates a decrease in randomness at solid/liquid interface at the time of the nickel adsorption by BF slag surface [76, 77].

Fig. 18
figure 18

Van’t Hoff plot for nickel adsorption by BF slag

Full size image
Table 5 Thermodynamic parameters of the nickel adsorption by BF slag
Full size table

Conclusions

In this work, the nickel adsorption by BF slag from aqueous solution was investigated as a function of contact time, stirring speed, initial pH, temperature, particle size, and initial concentration in batch adsorption technique. The physico-chemical characterization showed that the adsorbent consists mainly of the silica, lime, and alumina. The specific surface area of the BF slag grains is of the order of 275.8 m2/g. The optimization of the influencing parameters showed that the nickel removal efficiency and its maximum adsorption capacity, respectively, were 92.26% and 53.58 mg/g after 90 min of agitation. The kinetic studies showed that the mechanism of nickel adsorption process is well fitted by the pseudo-second order, the modeling results of kinetics indicated that the Langmuir model is more appropriate, and intra-particle diffusion is not the sole mechanism involved in this process. Thermodynamic study showed that the Ni(II) elimination by BF slag process is spontaneous, exothermic, and less entropic.

In conclusion, the results indicate that the BF slag represents a good adsorbent which can be used in the field of wastewater pollution control. Indeed, BF slag is better than the pyrophyllite, modified montmorillonite, raw red mud from aluminum industry, diatomite waste modified by EDTA, red mud, sea shells of Mehdia, and carbon aerogels [72, 78,79,80,81,82,83].