Evaluation of using electric arc furnace slag as an adsorbent for dyes removal

ABSTRACT The adsorption of two common textile dyes – orange G and methyl violet – onto an electric arc furnace slag was studied. Batch study shows that adsorption of orange G and methyl violet were dependent highly on pH value. Removal of orange G decreased with the increase of initial pH with optimum pH observed at 3. Moreover, the total removal percentage was low and did not exceed 45% at optimum pH. However, the removal of methyl violet increased with initial pH, and maximum removal (about 94%) was observed at pH 11. Based on the obtained data, methyl violet dye was chosen to study the effect of agitation speed, initial dye concentration, adsorbent dosage and adsorbent particle size on adsorption capacity. For the adsorbent under investigation, equilibrium adsorption capacity qe (mg/g) for orange G and methyl violet dyes were 17.4 mg/g and 3.29 mg/g, respectively. Also, batch studies show that the adsorption rate for methyl violet was rapid at the beginning and slowed down till equilibrium.In contrast, the orange G adsorption rate was slow until it approached equilibrium. The Pseudo-Second order model well represented the kinetics studies data. The value of K2 increased with adsorbent dosage while the K2 value decreased with an increase in initial dye concentration and adsorbent particle size. Equilibrium data were evaluated using Langmuir and Freundlich isotherm models and it was found that the Langmuir isotherm model best fitted to the experiments data.


Introduction
Textile processing and manufacturing expanded rapidly to fulfil the demand for such products.Dyes are a large and important class of that industry; they are widely used in different kinds of industries including textile, leather, and printing.Discharging wastewater that contains dyes, even in small amounts, could cause numerous severe issues to the receiving water bodies such as reducing the varieties of aquatic life by interrupting the penetration of sunlight which will cause in changing the behaviour of the dominant aerobic species and replaced by anaerobic species [1].Also, many by-products from chemical reactions will be present in the wastewater and most of these by-products are toxic and have carcinogenic properties [2].
Effluent from the textile process will contain many chemicals, mainly dyes and heavy metals, oil and grease, and suspended solids [3].Wastewater from dyes processing and manufacture has been reported to be challenging to be treated by standard treatment processes.Because most dyes are synthetic, obstinate in nature, and have a complex structure [4].Many processes have been used to treat dyes wastewater from different industrial processes.These processes include biological, physical and chemical methods [5].Examples of treatment methods include biodegradation [6], membrane filtration [7], chemical oxidation [8].However, many of these processes are unsuitable to be used as a treatment method for large wastewater.Hence, the adsorption process is implicit in being used as an industrial wastewater treatment option due to many considerationsfirst, its efficiency in pollutants removal, and secondly, its economic feasibility [9].Activated carbon considers as one of the essential adsorbents that have been used for this purpose, but it is processing and preparation is highly cost.Activated carbon produced from agricultural waste has become an essential source because of its cheap production cost and adequacy to treat wastewater from different industries [10].Several studies initialised to use different materials as a source for activated carbon.Jumasiah et al. studied the use of palm kernel shells as raw material for activated carbon for adsorption of basic dye.Redlich-Peterson isotherm equilibrium and Pseudo-secondorder adsorption kinetic are reasonably fit the results [11].
Electric arc furnace slag (EAFS) is produced during the steel manufacturing process [21].The amount of steel slag is increasing rapidly due to the increase of demand for steel products and an almost large amount of which will end up in disposal sites and becomes an environmental problem.Metal slag has been used in different treatment approaches such as treatment of an aqueous solution containing metals [22], phosphorus removal [23,24], heavy metals [25].Chemical analysis of the electric arc furnace slag shows a good consistency of the principal oxides.The electric arc furnace slag contains CaO, SiO 2 , FeO, MnO, MgO and Al 2 O 3 in high percentages [26], which increase the chances of using as an adsorbent of wastewater and may replace activated carbon.
However, no study has been reported on removing Orange G and Methyl Violet dyes using Electric Arc Furnace Slag.In this study, we attempt to utilise electric arc furnace slag (EAFS), a steel-making industry by-product, to remove Orange G and Methyl Violet dyes from synthetic wastewater.

Adsorbent and adsorbates
Electric arc furnace slag (EAFS) was used as an adsorbent without any pre-treatment.Slag (Figure 1) was provided from Amsteel Mills Sdn Bhd, a member of The Lion Group commenced.It operates two steel mills, in Klang and Banting, both in Selangor, Malaysia.EAFS was ground and sieved for different particle sizes in the range of 0.5 mm, 1 mm, 2 mm, 3 mm, and 4 mm.
Dyes used as adsorbates throughout the experiments were Orange-G (OG) and Methyl Violet (MV) from R & M chemicals, UK.Orange-G is categorised as a synthetic azo dye, while Methyl Violet is a dark green powder base dye.The characteristics and chemical structures of these dyes are shown in Table 1 and Figure 2.

Batch adsorption experiments
A dye stock solution with a concentration of 1000 ppm was prepared by dissolving 1.0 g of both Methyl Violet and Orange G dyes in 1000 mL distilled water.Each stock solution was shaken for 3 hours using an incubator shaker to ensure the complete dissolving of the dyes.Since methyl violet dye is difficult to dissolve in water, the dye solution was allowed to stand for one day until the absorbance of the solids and solution remained unchanged.In each experiment, the desired dye concentration was obtained by further dilution.A batch experiment was conducted to study the performance of EAFS in dyes removal.The effect of different initial dyes concentrations (20, 50, 80, 100, and 120 mg/L) was investigated.Different agitation speeds of 60, 80,  100, 120, and 150 rpm were tested.The time for each test was fixed to 30 minutes.Mass of adsorbent (EAFS) was studied in three values: 1 g, 3 g and 5 g.The effect of adsorbent particle size was also examined using three different sizes of EAFS: 0.5, 1.5, and 3 mm.The pH of the solution was studied in the range of 3 to 11 and was adjusted using 0.1 M HCl and 0.1 M NaOH.The effect of contact time was studied as follows: for the first (15 minutes), samples were withdrawn each (3 minutes) after that, samples were drawn each (5 minutes) for the next (15 minutes) followed by drawing each (10 minutes) for the next (60 minutes), for the following (2 hours), a (30 minutes) samples were drawn and finally (2 samples) with an interval of (1 hour).

Characterisation Methods
Fourier Transform Infrared Spectroscopy (FTIR) analysis was carried out using Thermo Nicolet Model Nicolet 6700 FTIR spectrophotometer.Images of scanning electron microscopy (SEM) of representative particle samples for both before and after the adsorption process were obtained using Hitachi Model S-3400 N PC-Based Variable Pressure.Analyses were undertaken to determine the surface heterogeneity of the electric arc furnace slag.Specific area was determined using nitrogen adsorption at liquid nitrogen temperature.The method used to achieve measurement is Brunauer, Emmett, and Teller (BET).To measure the surface characteristics, a Micromeritics ASAP BET instrument (GA, USA) was employed in these experiments.Nitrogen and helium were used as the adsorbate gas and carrier gas, respectively.The BET tube was loaded with between 0.65 and 1.0 g of the sample.Dyes concentrations were measured using UV/Vis double beam spectrophotometer (Dynamica Halo DB-20).The λmax for Methyl Violet (MV) and Orange G (OG) were found to be 584 nm and 492 nm, respectively.The pH measurements were conducted using the pH metre model pH1000 (Fisher Science).

Characterisation of adsorbent
The FTIR spectra of the adsorbent Figure 3 were recorded before and after adsorption in the range of 400-4000 cm-1 to determine the main groups of electric arc furnace slag that may participate in dye adsorption.From the displayed numbers of adsorption peaks, adsorption peaks of around 3643 cm-1 and 3441 cm-1 are attributed to the presence of hydrated minerals such as Ca(OH)2.The peak of around 1480 cm-1 attributed to carbonate anions CO32-ions.The peaks of 1029 cm-1 and 876 cm-1 are the Si-O and Al-O bonds of the aluminosilicate structure.
From Figure 3, it can be observed that there is a shift in the functional groups' peaks in the range of 670 cm-1 to 1480 cm-1 after the adsorption process, which indicates that these functional groups are likely to participate in the metal binding [27,28].
The surface image and morphology of the adsorbent electric arc furnace slag samples before and after adsorption were examined by scanning the electron microscopic (SEM) as shown in Figure 4.The SEM images show changes in the surface structure after the  sorption process.These surface mineralogy changes may result from dye molecules dispersal on electric arc furnace slag surface due to physisorption, the sorption of dye molecules to Si-specific sites, or a combination of these factors.
Results of the BET surface area of electric arc furnace slag samples are shown in Table 2. BET analysis was conducted to evaluate changes to surface area characteristics before and after experimentation and to determine the impacts of dye adsorption on the physical characteristics of electric arc furnace slag samples.
The post-treatment results show a decrease in surface area for electric arc furnace slag (0.1129 m 2 /g) as compared with that of the fresh slag (0.3249 m 2 /g), and likely indicates the dispersal of dye on the slag surface.Therefore, physical adsorption (or physisorption) appears to have contributed to dye removal.The pore size of the adsorbent particles increased after the adsorption process.This is most likely because the accumulation of the adsorbate (dyes) inside the adsorbent particles increases the pore size of the adsorbent.Also, the dyes molecules that are adsorbed on the active sites of the slag adsorbent might remove some of these active sites, leading to the reduction in the pore size of the adsorbent.

Effect of pH Value and Point of Zero Charge
The point of zero charges (pH pzc ) is the pH value where positive and negative charges are equal on the surface of a material, which makes it possible to describe the properties of the resulting interfaces.The surface charge of the adsorbent depends on the pH of the solution and pH pzc .For pH values lower than pH pzc , the adsorbent presents a positive surface charge that favours the adsorption of negatively charged substances, such as anionic dyes.On the other hand, when the pH is more significant than pH pzc , the adsorbent shows a negative surface charge that favours the adsorption of positively charged substances, such as methyl violet.The pH pzc of electric arc furnace slag was investigated by adding a calculated amount of slag to a solution of known pH.The solution was left for 24 hours.Then, the pH of the solution was determined and the difference in pH was calculated.Figure 5 shows that the point zero charges (pH pzc ) of electric arc furnace slag were found to be at pH 7. 4.
In order to study the effect of pH, the adsorption of dyes on metal slag was investigated in pH ranging 2 to 11 as shown in Figure 6.It can be seen that the maximum amount adsorbed for orange G (OG) was achieved at pH 3.However, the removal percentage was too low about 45%.In contrast, methyl violet adsorption starts at a pH higher than pHpzc (at pH 7, the removal percentage was 57%).The amount of adsorption capacity of methyl violet increased as pH increased.The removal percentage reached 93% at pH 11.At alkaline pH (pH more than pHpzc), the slag surface has a negative charge that leads to the increased attraction of positively charged dyes.Therefore, pH value 11 was selected for subsequent adsorption studies of methyl violet.

Effect of contact time
Contact time is an important parameter that can help to determine the efficiency of the adsorption process, equilibrium time, and the adsorption kinetics between adsorbent and dye molecules during the adsorption process.Methyl Violet (MV) adsorption was found to increase rapidly with time and fast on attaining equilibrium.On the other hand, Orange G (OG) adsorption was very slow and it takes more time to attain equilibrium.Adsorptions of Methyl violet and orange G with time onto electric arc furnace slag (EAFS) are shown below in Figure 7.
The initial concentrations of both dyes were fixed to 50 mg/L with optimum pH.The amount of methyl violet dye removed, as shown in Figure 7, was rapid in the first 15 minutes and reached 92.4%.In the first 3 to 15 minutes, the methyl violet dye absorbed increased from 40.4 mg/g to 46.2 mg/g.however, after 30 minutes, the amount of dye absorbed was almost constant, and the maximum removal percentage was reached 93% in which the amount of dye absorbed was 46.5 mg/g.On the other hand, Orange G (OG) removal percentage as shown in Figure 7, was low.In the first 60 minutes, the removal percentage did not exceed 40% in which the amount of dye absorbed increased from 7.3 mg/g at time 3 minutes to 19.7 mg/g at time 55 minutes.The maximum removal percentage of orange G was 45% at 85 minutes, in which the amount of dye absorbed was 22.9 mg/g.After 85 minutes equilibrium was reached and no changes in orange G removal were noticed.The rapid methyl violet dye adsorption in the first 3 to 20 minutes can be attributed to the increased availability of adsorption sites on the surface.After 30 minutes, methyl violet dye adsorption was slower, implying equilibrium conditions had been attained.

Effect of agitation speed
The effect of agitation speed on the adsorption rate of methyl violet was examined at constant temperature 28C° (+2), pH 11, 1 g of adsorbent, and 50 mg/L initial dye concentration.Agitation speeds under investigation were 60, 80, 100, 120, and 150 rpm.Agitation speed is an essential parameter for adsorption phenomena because it affects mass transfer processes and affects the outer boundary layer.Figure 8 shows the effect of agitation speed on methyl violet adsorption rate.As shown in Figure 9, the adsorption rate increased with the increase in agitation speed.When the agitation speed increased from 60 rpm to 120 rpm, the adsorption capacity increased from 45.78 mg/g to 46.54 mg/g, respectively.On the other hand, when agitation speed increased from 120 to 150 rpm, adsorption capacity decreased from 46.54 mg/g to 46.06 mg/g.From figure S1, percentage removal increased slightly up to 120 rpm.With a further increase in the agitation speed from 120 to 150 rpm, the adsorption capacity was found to be decreased.Therefore, the optimal mixing speed is around 120 rpm.The possible reason for the gradual increase in removal percentage with agitation speed can be related to more rigorous contact opportunities between dye ions and electric arc furnace slag surface.As a result, mixing speed can be considered an effective parameter in batch adsorption studies.

Effect of initial dye concentration
The influence of initial methyl violet dye concentration on the adsorption process was investigated at fixed adsorbent dosage (1 g), pH of 11 and Temp 28 C o (+2).Different dye concentrations (20, 50, 80, 100, 120 mg/L) and different time intervals (every 3 minutes for the first 15 minutes, then 5 minutes for the next 45 minutes, followed by 10 minutes interval for the next 60 minutes).As shown in Figure S2, increasing the initial concentration from 20 to 120 mg/L increased the adsorption capacity from 2.59 to 17.4 mg/g.This could be related to the domination of initial concentration mass transfer driving force over the mass transfer resistance between solution and solid phase.Therefore an increase in initial dye concentration increases the adsorbent adsorption capacity. Figure 9 shows that the equilibrium for low Methyl violet concentrations (20, 50 mg/L) was reached within the first 15 minutes.While a more extended time was needed for higher concentrations to reach equilibrium, an initial concentration of (120 mg/L) needed about 35 minutes to reach equilibrium.
The fast adsorption rate for low initial concentration can indicate that adsorption phenomena happened mainly on the surface of the adsorbent.In contrast, the high initial concentration adsorption process shows a rapid adsorption rate at first due to the availability of binding sites for dye.Then, adsorption starts to slowdown till it reaches equilibrium which indicates that the active site is fully occupied.

Effect of adsorbent dosage
The study of adsorbent mass effect on adsorption kinetic was performed using different adsorbent dosages in the range of (1 to 5 g), and an initial dye concentration of 50 mg/L.Figure S3 summarises the result in terms of adsorption rate versus adsorbent dosage.The result shows that as the adsorbent dose increased from 1 g to 5 g, the amount absorbed also slightly increased from 0.46 to 0.47 mg/g.The positive correlation between dye adsorption rate and adsorbent mass could be related to an increase in surface area and adsorption active sites [29,30].In Figure 10, plots of dye uptake (qt) and time, the results indicated that in all adsorbent dosages, the adsorption is rapid at first, slowing down towards equilibrium.At the initial stages of the adsorption process, a higher adsorbent mass shows a higher adsorption rate due to the high availability of adsorption-free sites that dye molecules can occupy.In general, increasing the adsorbent dosage did not show a significant improvement in the removal percentage; the 1 g dose can give the optimum result from both economic and performance prospect.

Effect of particle size
Three particle sizes of electric arc furnace slag were used to study the adsorption of methyl violet from an aqueous solution.Particle size less than (1.0 mm) in diameter consider small, while particle size larger than 1.0 mm and smaller than 3.0 mm consider medium and particle size larger than 3.0 mm consider significant.Figure S4 shows the amount of dye adsorbed versus particle size and Figure 11 illustrates the effect of contact time and particle size on adsorption equilibrium.The result indicates that smaller particle sizes have higher adsorption capacity.This can be due to a higher surface area-to-mass ratio at the smaller particle size available for adsorption.As seen in Figure S4, the adsorption capacity of small particle size (pz ≤ 1.0 mm) was 2.327 mg/g while medium particle size (1 mm < pz ≤ 3 mm) was 2.257 mg/g and large particle size (pz ≥ 4 mm) was 2.141 mg/g.Figure 11, indicates that the time for equilibrium was longer for large particle size and the adsorption rate was slow and gradual.At the same time, smaller particle size shows a high adsorption rate at the first stages of the adsorption process and then slows down until it reaches equilibrium.The result shows that a large particle size needs 80 minutes to reach equilibrium.In comparison, a small particle size needs 40 minutes to reach equilibrium.Furthermore, in terms of adsorption capacity, a small particle size needs 15 minutes to reach an adsorption capacity of 2.1 mg/g.while medium particle size needs 40 minutes and large particle size needs 70 minutes to reach the same adsorption capacity.

Adsorption kinetics
The kinetic data of methyl violet adsorption onto electric arc furnace slag was analysed using two models first-order and second-order kinetic models at temperature 28C°(±2).
The first-order kinetics was discussed in chapter two, in which the linear form is given as The plots of log (qe-qt) versus time are presented in Figures (12, 13, and 14) which indicate the applicability of using the first-order kinetic model to experimental data.The model constants K L , qe were obtained from slop and point of intercept and the data are shown in tables S1 to S3.
In spite of the values of correlation coefficient (R2) being considered to be high (0.8320 to 0.9047), the values of qe calculated were too low compared with the experiment value.
The effect of adsorbent mass shows that the value of KL increased with increasing in adsorbent mass due to an increase in available surface area.The value of qe calculated did not match the experiment value.The effect of particle size on value KL shows that by increasing the adsorbent particle size, the value of KL decreased.Also, the values of qe calculated are too low compared with the experiment value.
The linear form of the second-order kinetic is given as: A straight line can be obtained from the plots of t/qt versus time.Figures 15,16  From Table S4 to S6, we can see that the values of qe calculated match the experiment values.Also, the values of the correlation coefficient (R 2 ) are high in the range of (0.9992 to 0.9999).
Table S4 shows that the values K2 decreased with the increase of initial concentration at a range of 0.0449 at 20 mg/L to 0.0032 at 120 mg/L.The same effect was observed with a particle size as shown in Table S5, where the value of K2 changes from 0.0113 for small particles size to 0.0026 for small particles.In contrast, the value of K2 increased with an increase in adsorbent mass as seen in Table S6 where the value of K2 changed from 0.0035 at 1 g to 0.012 at 5 g dosage.
As shown, the values of qe calculated from pseudo-second-order are in match with experiment values.The value of the correlation coefficient for the pseudoseconder order model is higher than for the pseudo-first-order model; these results indicate the rate of adsorption is following the pseudo-second-order model.

Adsorption isotherms
The adsorption equilibrium studied was performed using two of the most common adsorption isotherms, Langmuir and Freundlich.
Langmuir isotherm is used to investigate the adsorption process occurring in monolayer.The linear form of the Langmuir isotherm is shown below  Plotting q e C e versus qe gives the Langmuir monolayer adsorption constant / L which is represented by the intercept and adsorption energy constant KL represented by the slop.Figure 18 show the applicability of Langmuir isotherm to experiment data.While equilibrium adsorption parameters KL, αL, and RL are presented in Table S7.Langmuir isotherm can be expressed in dimensionless separation factor (RL) which predicts whether the adsorption system is favourable or unfavourable.The following relationship can define separation factor: Where RL is the separation factor, KL is Langmuir constant and C o is the initial or reference concentration.The value of RL indicates the isotherm shape according to Table S8.
The Freundlich isotherm assumes that dye uptake occurs on the heterogeneous surface by multilayer adsorption and by increasing the adsorbate concentration, the amount adsorbed will increase infinitely.The linear form of the Freundlich equation is shown below: A plot of ln qe versus ln Ce indicates the value of K F and n from intercept and slope, respectively.A higher value of K F indicates that the uptake of dye is easy.While the value n between 1 and 10 shows a favourable adsorption process.Figure 19 shows experiment data modelled by the Freundlich isotherm model, and Table S9 shows isotherm parameters.
The value obtained for both Langmuir and Freundlich, as summarised in Table (S7) and Table (S9), shows that the value of the correlation coefficient (R 2 ) is 0.991 and 0.9893, respectively.The higher value of the correlation coefficient of the Langmuir isotherm model indicated that the system fits the Langmuir isotherm very well.At the same instance, the dimensionless parameter RL values 0.037.This RL value between zero and unity indicates that Methyl violet dye adsorption is favourable to Langmuir isotherm throughout the 180 minutes adsorption study time.

Conclusion
This study has shown the applicability of using electric arc furnace slag as an effective adsorbent for orange G and methyl violet.Electric arc furnace slag was found to be an effective adsorbent for methyl violet; the maximum removal percentage observed was 93%.However, electric arc furnace slag shows a low attitude towards removal of orange G.It can be seen that the maximum amount adsorbed for orange G was meagre, about 45%.Particle size experiments showed that smaller particle sizes had a higher adsorption rate at the first stages of adsorption than large particle sizes and then slowed down until they reached equilibrium.While large particle size equilibrium time was long and adsorption rate was slow and gradual.The result indicates a positive correlation between dye adsorption rate and adsorbent mass.The adsorption process for both orange G and methyl violet was highly dependent on pH.Removal of orange G was observed at a low pH value of 3.0 and decreased with an increase in pH.However, the removal of methyl violet was achieved at high pH.During the experiment process, increasing pH value increased sufficiently the removal rate and the optimal pH observed at pH 11.
The Langmuir isotherm showed the best agreement for the experiment data in terms of the correlation coefficient.Therefore, the adsorption of methyl violet onto electric arc furnace slag occurs through monolayer formation on the adsorbent.Also, Langmuir isotherm dimensionless separation factor (RL) value showed that the adsorption system is favourable.
Batch kinetic studies of methyl violet adsorption indicate the adsorption was rapid at the beginning and slowed down until approaching equilibrium.Experiment data modelled using pseud-first order and pseudo-second-order.The best-fitting model was found to be pseudo-second-order in a whole range of adsorption processes.Hence, the rate of dye removal from aqueous solution was dependent on electrostatic interactions between dye molecules and adsorbent surfaces.

Figure 2 .
Figure 2. Molecular structures of the Orange-G and methyl violet dyes.

Figure 6 .
Figure 6.pH effect on removal percentage of methyl violet and Orange G.

Figure 7 .
Figure 7. Contact time and removal percentage of Methyl Violet and Orange G.

Figure 8 .
Figure 8. Agitation speed effect on adsorption rate of Methyl Violet.

Figure 9 .
Figure 9. Contact time and initial concentration effect on Methyl Violet adsorption rate.

Figure 10 .
Figure 10.Contact time and adsorbent dosage effect on adsorption rate.

Figure 11 .
Figure 11.Effect of contact time and particle size on adsorption rate.
and 17   indicate the applicability of the second-order kinetic model to experimental data.The model constants K2, qe were obtained from slop and point of intercept, and the data are shown in tables S4 to S6.

Figure 12 .
Figure 12.First-order kinetic model: effect of initial concentration.

Figure 14 .
Figure 14.First-order kinetic model: effect of particle size.

Figure 15 .
Figure 15.Second-order kinetic model: effect of initial concentration.

Figure 17 .
Figure 17.Second-order kinetic model: effect of particle size.

Table 1 .
Characteristics of the Orange-G and methyl violet dyes (source R&M Chemicals).

Table 2 .
BET surface area of electric arc furnace slag.