1 Introduction

During the past decades, the urgent situation with scarcity of domestic crude oil and natural gas as well as relative abundance of coal resource has significantly boosted the development of coal-based chemical industries in China [1]. However, these industries inevitably produce massive coking wastewaters during coal gasification, coal gas purification and by-product recovery processes [2, 3]. In general, the coking wastewater contains plenty of non-biodegradable and refractory organic pollutants that are highly concentrated, toxic and carcinogenic [4]. Direct discharge of these coking wastewaters can cause serious natural water pollution and may have long-term environmental and ecological impacts.

Quinoline is a typical N-heterocyclic aromatic compound that is present in the coal gasification wastewater. Nowadays, it remains challenging to achieve effective removal of quinoline from wastewater due to their relatively high aqueous solubility and low biodegradability [2, 5]. Up to date, numerous investigations have been undertaken on the withdrawal of quinoline from wastewater via a variety of biological, physical and chemical methods, such as biological treatment [6], adsorption [7], wet oxidation and catalytic wet oxidation [8, 9], Fenton-like [10], photocatalysis [5], and ozonation [11] etc. Largely owing to its simplicity and widely available adsorbent materials, adsorption has been considered as one of the most promising alternatives to the treatment of organic pollutants in aqueous solutions. Activated carbon has long been regarded as an ideal adsorbent. However, the high cost of activated carbon has stimulated the interest in replacing it with various solid waste materials [12].

The utilization of coal also produces a solid waste, coal fly ash (CFA), which is a major by-product of thermal power plants all over the world [13]. The amount of annually discharged CFA is enormous and its improper disposal has become a serious environmental concern [14]. From the perspective of physicochemical characteristics, CFA is favorable for use as an adsorbent for treating various organic pollutants from an aqueous solution [15]. During the past decades, many researchers have evaluated the performance of CFA as cheap adsorbent for handling various organic contaminants including phenolic compounds [16, 17], dyes [18, 19], lignin [20], tannin [21], ciprofloxacin [22], etc. However, the adsorption performance strongly depends on the activation methods of CFA [4, 18]. These activation processes usually involve chemical treatments of CFA with strong acids or bases that can inevitably lead to a secondary pollution, making it more attractive to use raw CFA directly. Until now, the scientific researches regarding direct utilization of raw CFA for treating coal gasification wastewater are very rare, and more detailed evaluations are needed for better illustrating the intrinsic adsorption properties of raw CFA for quinoline.

In this work, we investigate the intrinsic adsorption properties of the as-received CFA for quinoline from aqueous solution. The dependence of adsorption on CFA dosage, initial quinoline concentration and contact time were systematically studied in batch mode. We adopted the most widely used adsorption isotherms and kinetic models to fit the experimental data and discussed the potential adsorption mechanism.

2 Experimental

2.1 Material and characterization

The CFA was obtained from Qingdao Thermal Power Plant, China. The CFA was used without further treatment. Quinoline was purchased from Tianjin Beilian Fine Chemicals Development Co., Ltd., China. The composition of the raw CFA was determined by X-ray fluorescence spectrometry (XRF) in a Bruker S4 Pioneer spectrophotometer. The morphology of the CFA was measured with scanning electron microscopy (SEM, JSM 6700 F, operating at 8 kV). Powder X-ray diffraction (XRD) pattern of the fly ash was recorded on a D/max-2500/PC X-ray diffractometer (Rigaku, Japan) with a Cu Kα radiation source. The N2 adsorption–desorption isotherms were carried out at 77 K on a surface area analyzer (Micromeritics, ASAP2020, USA). The surface area was estimated by the Brunauer–Emmett–Teller (BET) method.

2.2 Adsorption experiments

Adsorption of quinoline on CFA was conducted by directly mixing the CFA with standard quinoline solution. In a typical adsorption experiment, 0.25 g of CFA was added to 50 mL of quinoline solution with initial concentration of 50 mg/L and maintained for the time corresponding to the analysis. All the adsorption experiments were conducted at the initial pH of quinoline solution (7 ± 0.2). Then, the CFA was filtered with a 0.45 μm polytetrafluoroethylene syringe membrane filter and the filtrate solution was analyzed for obtaining the remaining concentration of quinoline. The kinetic experiments were carried out by varying the initial quinoline concentration (20–100 mg/L), solution temperature (25–45 °C) and adsorbent dosage (2–10 g/L), respectively. All of the tests were carried out in duplicate. The analysis of quinoline concentration was performed by using an UV–vis spectrophotometer (Yoke).

The equilibrium adsorption capacity (qe) and time-dependent adsorption capacity (qt) are calculated from the following equations, respectively:

$$q_{e} = \frac{{V\left( {C_{0} - C_{e} } \right)}}{m}$$
(1)
$$q_{t} = \frac{{V\left( {C_{0} - C_{t} } \right)}}{m}$$
(2)

where C0, Ct and Ce are the initial, time-dependent and equilibrium quinoline concentrations (mg/L), respectively. V (L) is the volume of solution and m (g) is the adsorbent amount.

3 Results and discussion

3.1 Characterization of CFA

The properties of the CFA sample are as shown in Table 1. SiO2 and Al2O3 are the main components with a total content of 58.8 wt%. The specific surface area and pore volume of the CFA obtained from the N2 equilibrium adsorption isotherms are 8.84 m2/g and 0.016 cm3/g, respectively, as shown in Table 2. The XRD pattern of the as-received CFA sample is presented in Fig. 1. The major crystalline phases were identified to be quartz (SiO2), mullite (3Al2O3 SiO2), hematite (Fe2O3), and anorthite (CaAl2Si2O8), which are in agreement with those reported by Mishra [23] and Yilmaz [24]. In addition, amorphous silicate phase was also detected as a broad reflection at 20°–30° in the XRD pattern [25]. The SEM images of the as-received CFA at different magnifications are presented in Fig. 2. The CFA comprises typical glassy spherical-shaped particles as shown in Fig. 2a. It might also be observed that some smaller size particles with various shapes are adhered to bigger size particles (Fig. 2b).

Table 1 Chemical compositions of coal fly ash (wt%)
Full size table
Table 2 Pore structure data of coal fly ash
Full size table
Fig. 1
figure 1

XRD pattern of the raw coal fly ash

Full size image
Fig. 2
figure 2

SEM images of the raw coal fly ash at a × 1000 and b × 40,000

Full size image

3.2 Effect of CFA dosage, initial quinoline concentration and contact time

The influence of adsorbent dosage on the adsorption of quinoline is given in Fig. 3. In case of 0.10 g CFA dosage, an adsorption capacity of 7.46 mg/g was acquired, which is comparable with the maximum capacity of 2.1 mg/g for coking wastewater adsorption [26] and 3.95 mg/g for Reactive Black 5 adsorption on raw coal fly ash [27]. It can be noticed that the amount of quinoline adsorbed per mass of adsorbent by CFA markedly decreases from 7.46 to 4.95 mg/g with the increment of adsorbent amount in the range of 0.1–0.4 g. Accordingly, the removal efficiency increases with an increase in adsorbent dosage due to the unsaturation of adsorption sites at higher adsorbent dosages for constant quinoline concentration (not shown). When the amount of CFA is further increased to 0.5 g, the adsorption capacity of CFA for quinoline is slightly reduced to 4.50 mg/g, meaning the slow increase of quinoline removal as both the surface and concentration of quinoline settled to equilibrium with each other [28].

Fig. 3
figure 3

Effect of adsorbent dosage on adsorption of quinoline onto CFA at 25 °C (initial quinoline concentration: 50 mg/L)

Full size image

The influence of contact time on quinoline adsorption onto CFA at different initial quinoline concentrations (20, 50, and 100 mg/L) is displayed in Fig. 4. It was observed that the adsorption amount of quinoline increases with the increment of contact time. Particularly, a rapid adsorption occurred during the first 20 min irrespective of the initial quinoline concentrations, which is most possibly ascribed to the high concentration gradient and lots of vacant sites available during the initial adsorption stage. Since the majority of the exterior sites of the adsorbent surface were occupied with quinoline molecules after the fast adsorption stage, the adsorption began to occur at internal sites of CFA and proceeded gradually until a final equilibrium realized after about 140 min. At the same time, the continuous reduction in the concentration driving force with the prolonging of contact time also retards the adsorption. Similar result was also reported for the adsorption of quinoline onto different adsorbents in the literature [29]. On the other hand, the adsorption capacity is closely related to the initial quinoline concentration and exhibits a marked increase with the concentration due to higher accessibility of quinoline molecules in the solution for adsorption and concentration gradient for mass transfer between the aqueous phase and the solid phase.

Fig. 4
figure 4

Effect of contact time and initial concentration on adsorption of quinoline onto CFA at 25 °C (catalyst dosage: 0.25 g)

Full size image

3.3 Adsorption kinetics

In this study, three simplified kinetic models (pseudo-first-order, pseudo-second-order and intra-particle diffusion) were applied to correlate the experimental data. The pseudo-first-order (expressed by Eq. (3)) and pseudo-second-order (expressed by Eq. (4)) kinetic models assume that the physical adsorption and chemical adsorption dominate the rate, respectively.

$$\ln (q_{e} - q_{t} ) = \ln q_{e} - k_{1} \cdot t$$
(3)
$$\frac{t}{{q_{t} }} = \frac{1}{{k_{2} q_{e}^{2} }} + \frac{1}{{q_{e} }}t$$
(4)

in which k1 (min‒1) and k2 (g/mg min) are the rate constants, and qt and qe represent the amounts of quinoline adsorbed (mg/g) on adsorbent at the sampling time t (min) and at equilibrium, respectively.

The fitted lines of ln(qe − qt) against t for pseudo-first-order and t/qt versus t for pseudo-second-order kinetic models are shown in Fig. 5a and Fig. 5b, respectively. All the parameters of qe and k as well as correlation coefficient constant R2 calculated from the corresponding fitted lines are summarized in Table 3. It is noticeable that the qe values calculated with the pseudo-second-order kinetic model are much more close to the experimental data. By comparing values of R2, it can be concluded that the experimental data are better fitted with the pseudo-second-order model, meaning that the adsorption process is likely to be a chemisorption process, which depended on the chemical character of CFA and quinoline solution. Similar finding with respect to the fit of pseudo-second-order was provided by researchers on quinoline adsorption onto bamboo charcoal [30], anionic dyes sorption on fly ash [31] Reactive Black 5 adsorption on high lime coal fly ash [27].

Fig. 5
figure 5

Linear fit of experimental data using a pseudo-first-order and b pseudo-second-order kinetic models

Full size image
Table 3 Parameters of pseudo-first-order and pseudo-second-order adsorption kinetic models at different initial quinoline concentrations
Full size table

The intra-particle diffusion model described by both external and intra-particle diffusions is usually adopted to explore the rate controlling steps involved during adsorption. The intra-particle diffusion model is expressed as follows:

$$q_{t} = k_{i} t^{1/2} + C$$
(5)

where ki is the intra-particle diffusion rate constant (mg/g min1/2) and the values of C depict the boundary thickness.

To complete the adsorption of quinoline onto CFA, the adsorbate should diffuse from the solution to the external surface and subsequently to the internal surface of the adsorbent (intra-particle diffusion). In our study, the regression of qt versus t1/2 shows a bi-linearity nature (Fig. 6), meaning two steps of mass transfer occur during the adsorption process [32, 33]. Obviously, this bi-linearity became prominent at higher quinoline concentration. From Table 4, the values of ki,1 and ki,2 (the slopes of the two linear portions, respectively) increase with the increment of initial quinoline concentration owing to increase of driving force at higher initial concentrations. The adsorption stage 1 is ascribed to the external mass transfer, corresponding to the easy adsorption of quinoline molecules on exterior sites of CFA. The relatively slow adsorption stage 2 is possibly on account of the low residual adsorbate concentrations after experiencing an external mass transfer process [34, 35] and/or the diffusion of adsorbate molecules into the pores of the adsorbent [36]. As the fitted lines do not pass through origin, it can be assumed that intra-particle diffusion is not the only rate-controlling step [37]. Many researchers have also reported that the adsorption rate is generally determined by both intra-particle diffusion and liquid film diffusion steps [38, 39].

Fig. 6
figure 6

Linear fit of experimental data using intra-particle diffusion model

Full size image
Table 4 Intra-particle diffusion rate constants of quinoline adsorption onto coal fly ash
Full size table

3.4 Adsorption isotherm

Langmuir, Freundlich and Redlich-Peterson adsorption isotherms are widely used for describing qualitative information about the capacity of various adsorbents and the nature of the solute-surface interactions [40,41,42]. The Langmuir isotherm assumes structurally homogeneous surface and monolayer adsorption system, which is described by the following equation:

$$q_{e} = \frac{{K_{L} q_{\hbox{max} } C_{e} }}{{1 + K_{L} C_{e} }}$$
(6)

where KL (L/mg) and qmax (mg/g) are the constants related to adsorption affinity of the binding sites and adsorption capacity.

The Freundlich model is based on the assumption of heterogeneous surfaces present in the adsorbents, which can describe heterogeneous surfaces and multilayer adsorption systems. The Freundlich isotherm is generally given as follows:

$$q_{e} = K_{F} C_{e}^{1/n}$$
(7)

where KF (mg/g) and n (dimensionless) represent the adsorption capacity and intensity.

The Redlich-Peterson model expressed as Eq. (8) includes three parameters and presents both the Langmuir and Freundlich isotherms.

$$q_{e} = \frac{{K_{RP} C_{e} }}{{1 + \alpha C_{e}^{\beta } }}$$
(8)

where KRP (L/g) and α (L/mg)β are Redlich-Peterson isotherm constants, β is the exponent which lies between 0 and 1.

Figure 7 shows the equilibrium data of quinoline adsorption onto CFA at different temperatures as well as the nonlinear fits of Langmuir, Freundlich, and Redlich–Peterson isotherm models. The values of isotherm parameters and the determination coefficient (R2) are listed in Table 5. Obviously, Langmuir isotherm has higher R2 values than Freundlich and Redlich-Peterson models. Thus, Langmuir model is more appropriate for describing the adsorption process, indicating the quinoline adsorption occurred at mono-layer and on the homogeneous surface of CFA. This result is similar with that in some previous studies. Hsu [43] reported that the adsorptions of Acid Red 1 onto the raw and modified coal fly ashes could well fit the Langmuir isotherm model.

Fig. 7
figure 7

Nonlinear fits of equilibrium data at different temperatures with a Langmuir, b Freundlich, and c Redlich–Peterson isotherm models

Full size image
Table 5 Isotherm parameters for the adsorption of quinoline onto coal fly ash
Full size table

4 Conclusions

The adsorption studies reveal that the adsorption of quinoline on CFA is affected by initial quinoline concentration, contact time and adsorbent dosage. Kinetics studies indicate that the adsorption complies well with pseudo-second-order and intra-diffusion models, suggesting that the adsorption of quinoline on CFA is a possible chemisorption. It was also found that the Langmuir model can well describe quinoline adsorption isotherm, thus the homogeneous adsorption was speculated to be the main process. Despite relatively low adsorption capacity of the studied fly ash, the use of this adsorbent for wastewater removal is of interest since it is a low-cost and readily available waste. However, to enhance the adsorption capacity of CFA toward quinoline, an effective activation method that does not generate additional pollutants needs to be developed in the future.