Adsorptive removal of tetracycline by sustainable ceramsite substrate from bentonite/red mud/pine sawdust

In this study, a novel, sustainable and efficient ceramsite substrate of constructed wetlands (CWs) were prepared for tetracycline (TC) removal by employing bentonite (Ben) and red mud (Rm) as the main materials and pine sawdust (Ps) as the additive. The optimal parameters for Ben/Rm/Ps ceramsite preparation were obtained via orthogonal and one-factor experimental designs, and the optimal parameters were presented as follows: mass ratio of Ben: Rm: Ps = 4:1:0.9, preheating temperature = 240 °C, preheating time = 20 min, calcining temperature = 1150 °C, and calcining time = 14 min. The properties of Ben/Rm/Ps-op ceramsite (obtained at the optimal condition) were first analyzed, including XRD and SEM, and demonstrated a microporous structure with some crystal strength components. Neutral condition and higher temperature were indicated conducive to improve the TC removal efficiency, while coexisting ions (Na+ or Ca2+) showed adverse effect for TC adsorption by Ben/Rm/Ps-op. In addition, adsorption kinetics and isotherm could be well described by the second-order kinetics and linear isothermal model, respectively, which suggested chemisorption and multilayer adsorption thickness increased infinitely. The theoretical maximum TC adsorption capacity of Ben/Rm/Ps-op at 20 °C reached up to 2.5602 mg/g. In addition, Ben/Rm/Ps-op could effectively remove TC as the CWs substrate under a dynamic flow condition. Further, Ben/Rm/Ps-op exhibited high reusability capability and stability for TC removal, and the adsorption amount still remained for 2.13 mg/g (C0 = 80 mg/L) after three consecutive cycles.

Static adsorption experiments. All the adsorption experiments were carried out in dark using brown glass vials (total volume = 150 mL) with 50 mL TC solution and ceramsite on an HZQ-120H heating oscillator (Yiheng Scientific Instrument Co., Ltd., Shanghai, China) with a speed of 160 rpm. pH was adjusted using dilute HCl and NaOH aqueous solution (aq.).
TC adsorption kinetics studies were conducted at pH = 7 and temperature = 20 °C with an initial concentration of TC of 80 mg/L and ceramsite dosage of 20 g/L. At predetermined times (5-600 min), the vials were sacrificially sampled. Besides, to investigate adsorption thermodynamics, the adsorption kinetics assays were carried out at 30 °C and 40 °C as well. For TC adsorption isotherm experiment, the initial concentration of TC was varied from 2 to 80 mg/L with a fixed ceramsite dosage of 20 g/L, and the mixture (pH = 7) was shaken for 24 h at 20 °C to reach the adsorption equilibrium. To explore effect of ceramsite dosage on adsorption, different doses of ceramsite (5-50 mg/L) were added into TC solution (80 mg/L), and the mixture (pH = 7) was shaken for 24 h at 20 °C. To probe effect of pH on adsorption, the equilibrium tests were conducted with an initial TC concentration of 80 mg/L, a ceramsite dosage of 20 g/L, and finial solution pH 2-10 at 20 °C for 24 h. To examine effect of ionic strength, 0-0.25 mol/L NaCl or CaCl 2 were added into TC solution (80 mg/L) with ceramsite dosage of 20 g/L at pH = 7, temperature of 20 °C and shaking for 24 h. After adsorption is completed, the solution was filtered through a 0.22 μm microfiltration membrane. The concentrations of TC were detected via SP-756P Ultraviolet-Visible Spectrophotometer at 355 nm.
The adsorption amount at predetermined time t (q t , mg/g) and equilibrium adsorption amount (q e , mg/g) of TC on materials and removal efficiency (R, %) were calculated via: where C t (mg/L) is the residual concentration in the liquid phase at sampling time t (min); C 0 and C e (mg/L) are the initial and equilibrium concentrations of TC, respectively; V (L) is the total volume of the solution; and m (g) is the mass of ceramsite. TC concentration in the solution phase (C d , mg/L) was determined upon centrifugation and filtration, and the percent of TC desorbed as calculated via: Dynamic column experiments. To better appraise the adsorption efficiency of Ben/Rm/Ps-op for TC, dynamic column experiments were carried out. A glass column wrapped in aluminium foil was employed in the tests, with a height of 50 cm and an internal diameter of 5 cm. Ceramsites were loaded into the column with a height of 20 cm, and the corresponding filter volume was 393 mL. The initial concentration of TC was set as 4 mg/L, and the solution was bumped into the column in an up-flow mode. For the effect of HRT, column tests were carried out at different HRTs ( and volatiles also present in Ben and Rm. About 10.48% and 50.68% of fluxing components are respectively contained in Ben and Rm. Therefore, Ben and Rm can be used as the main materials for firing ceramsite. Moreover, the mineral constituent of Ben and Rm were appraised by XRD. For Ben, the peak strength of SiO 2 crystal phase is most intense, and crystal peaks of CaO 3 can also be observed (Fig. 1a). Since other major components of Ben such as Al 2 O 3 and MgO did not form crystal morphology, corresponding peaks were not detected. For Rm, large amount of crystal SiO 2 and a certain amount of CaCO 3 and Ca 3 Al 2 O 6 were detected (Fig. 1b).
According to elemental compositions in Table 1, we calculated the content of SiO 2 , Al 2 O 3 , fluxing components and SiO 2 + Al 2 O 3 in the mixture of Ben and Rm with different ratios (1:1, 2:1, 3:1, 4:1, 5:1, 6:1 and 7:1) ( Table 2). With the increase of proportion (Ben: Rm), the content of SiO 2 and SiO 2 + Al 2 O 3 would increase, while the content of Al 2 O 3 and fluxing components would decrease. Taking into account Riley three-phase diagram 30 and the actual situation of adding as much Rm as possible to the raw materials, four ratios (3:1, 4:1, 5:1 and 6:1) were adopted to prepare Ben/Rm ceramsite.
To optimize the preparation of CFA/WS ceramsite and identify the critical factors of determining the ceramsite properties, an orthogonal experimental design of five factors (mass ratio of Ben: Rm, preheating temperature and time, calcining temperature and time) and four levels (L 16 (4) 5 ) were conducted and bulk density of ceramsite was employed as the evaluation index. Relatively lower bulk density is preferable, which demonstrates higher porous ceramsite bodies 31 . The significance levels of different influencing factors on the ceramsite bulk density were clarified through the range analysis 32 . Table 3 summed up the results of L 16 (4) 5 orthogonal design. The K value for each level of a parameter was the average of four bulk density values, and the range value (R) for each factor was the difference between the    www.nature.com/scientificreports www.nature.com/scientificreports/ maximal and minimal K values of the four levels. The range analysis suggested Ben: Rm ratio was the most important factor and followed by preheating time and calcining temperature, while others factors were not of significance.
As depicted in Table 3, with the increase of Ben: Rm ratio, the bulk density of ceramic displayed a decline trend, which may be related to the change of sintering and volatile compositions. In addition, when the preheating time was within 15-25 min, ceramsite bulk density increased obviously with the increase of time. However, increasing the preheating time from 25 to 30 min, the bulk density showed a slightly decrease. Under certain conditions, increasing preheating time facilitated softening ceramsite, which ensured ceramsite produced enough gas in the roasting stage, thus reduced the bulk density. However, when preheating time is too long, organic matter and carbonate will decompose and volatilize to produce gas in the preheating stage, which will reduce the gas amount produced in roasting stage, then the bulk density increased.
Preparation and optimization of Ben/Rm/Ps ceramsite. Ps was characterized by ultimate analysis (Table 4) and TG (Fig. 2). The TG curve shows that the Ps mass nearly did not change below 260 °C, while a drastic decline occurred within 260-468 °C, and above 468 °C little Ps was residual (Fig. 2). Therefore, decomposition of Ps is expected to occur in the initial stage of ceramic calcination stage, which can improve the sintering performance and promote the formation of ceramic porous structure.
The aforementioned L 16 (4) 5 orthogonal experimental design indicated the mass ratio of raw materials and preheating time were the most significant influence factors on ceramic sintering. Therefore, for the preparation of Ben/Rm/Ps ceramsite, the mass ratio of raw materials and preheating time were optimized by using one-factor experimental design, while the other fixed parameters were adopted the optimal ones based on L 16 (4) 5 results, i.e., preheating temperature = 240 °C, calcining temperature = 1150 °C, calcining time = 14 min.
For the optimization of mass ratio of raw materials, the preheating time was fixed at 15 min, and four levels of Ben: Rm were employed, i.e., 3 g:1 g, 4 g:1 g, 5 g:1 g and 6 g:1 g. Different amounts of Ps (0.1-1 g) were added to the raw materials to sinter ceramsite. Apparent density, bulk density and compressive strength of ceramsite were determined to appraise the influence of raw materials ratio (Fig. 3a-c). Besides, a three-phase diagram based on the three parameters was presented in Fig. 3d, during which 19, 36 and 8 represent mass ratios of Ben: Rm: Ps = 4:1:0.9, 6:1:0.6 and 3:1:0.8, respectively. Combining the three-phase diagram ( Fig. 3d) with apparent density, bulk density and compressive strength curves (Fig. 3a-c), 8, 19 and 36 displayed outstanding performance among all ratios. Besides, 19 and 36 have little difference in compressive strength and apparent density, while showed much more excellent than 8. In addition, comparing with 36, 19 exhibited smaller bulk density and possessed more red mud, which is conducive to waste utilization. Therefore, the Ben: Rm: Ps ratio of 4:1:0.9 was chosen for the follow-up experiment.    Table 4. Ultimate analysis data of pine sawdust (wt%).
For the optimization of preheating time, the optimal ratio of Ben: Rm: Ps (4:1:0.9) was employed and the preheating time was varied from 10 min to 30 min. TC adsorption capacity of obtained ceramsite was assessed by static adsorption experiments. The TC adsorption capacity of ceramsite showed a trend of firstly increasing and then decreasing with the increase of preheating time (Fig. 4). From 10 min to 20 min, the TC adsorption capacity of ceramsite showed a steady increase, and reached the maximum value at 20 min. With the increase of preheating time, softening degree of ceramsite raises, which is conducive to ensuring enough gas producing in  the roasting stage, thus reducing bulk density. Continuously increase the preheating time from 20 min to 30 min, the TC adsorption capacity of ceramsite displayed a quick decline. This can be explained that the organic matter and carbonate will decompose and volatilize to produce gas in the preheating stage for too long preheating time, which will reduce the amount of gas produced in the roasting stage and increase the stacking density.
To sum up, the optimal parameters for preparation of Ben/Rm/Ps ceramsite were determined as: Ben: Rm: Ps = 4:1:0.9, preheating temperature = 240 °C, preheating time = 20 min, calcining temperature = 1150 °C, and calcining time = 14 min. The Ben/Rm/Ps ceramsite prepared at this optimal condition (hereafter referred to as Ben/Rm/Ps-op) had an apparent density of 1.41 g/cm 3 , bulk density of 0.54 g/cm 3 and compressive strength of 19.45 MPa.
SEM analysis (Fig. 5) showed that both the surface and cross-section of Ben/Rm/Ps-op was mainly distributed by macroporous and microporous structures with different sizes, indicating a large specific surface area and widely distributed adsorption sites. The SEM results demonstrated that ceramsite was a porous ceramsite and might be a good adsorbent for wastewater treatment. XRD detected the formation of several crystal phases in Ben/Rm/Ps-op including SiO 2 , Ca 3 Si 2 O 7 , Ca 2 Si 2 O 5 (OH) 2 and Ca 2 SiO 4 (Fig. 6). The crystal compositions are helpful to improve the strength of ceramsite; the active components such as SiO 2 , Si 2 O 7 6− , Si 2 O 5 (OH) 2 4− and SiO 4 4− can be adsorption sites of TC via ion exchange. Since TC contain a positively charged group in the structure, regardless of the zero net charge or negative net charge, it is likely that the molecule arranges at the surface in such a way that the positively charged group is located very close to the surface. SiO 2 is negatively charged at the solution of pH > 2.5, which could easily combine with TC. In addition, negatively charged groups Si 2 O 7 6− , Si 2 O 5 (OH) 2 4− and SiO 4 4− also showed excellent binding ability with TC. Moreover, heavy metal leaching toxicity analysis (Table 5) revealed the concentrations of leached heavy metals from Ben/Rm/Ps-op were far below hazardous wastes standard (GB 5085.   29 , and basically met the surface water quality of Class III (GB 3838-2002) 28 , recommending Ben/Rm/Ps-op will not cause secondary pollution to the aquatic environment. Therefore, Ben/Rm/Ps-op is a safe ceramsite with great mechanical strength and adsorption capacity. Fig. 7 that Ben/Rm/Ps-op could catch most of TC from aqueous solution in the initial 180 min and reach equilibrium after 360 min. After calculation, 64% TC was removed at equilibrium and the equilibrium adsorption capacity was as high as 2.5602 mg/g.

Adsorption of TC. Adsorption kinetics. It is observed in
Pseudo-first-order and pseudo-second-order models were tested to analyze the kinetics results, which are expressed as 33,34 : Pseudo-first-order model: where q t and q e (mg/g) are the adsorption capacities of TC at time t (min) and equilibrium, respectively; k 1 (min −1 ) is the rate constant for pseudo-first-order model and k 2 (g/(mg•min)) is for pseudo-second-order model, respectively. As a result, the pseudo-second-order model showed higher R 2 = 0.9996, compared with R 2 = 0.9789 for the pseudo-first-order model ( Table 6, Fig. 8). Besides, q e,cal (2.6889 mg/g) of pseudo-second-order model was closer to q e,exp value (2.5602 mg/g) than q e,cal (1.7591 mg/g) of pseudo-first-order model did. Consequently, the adsorption of TC by Ben/Rm/Ps-op conforms to pseudo-second-order model, indicating that the rate controlling step for adsorption was a chemical interaction 35 . www.nature.com/scientificreports www.nature.com/scientificreports/ Adsorption mechanism. The pseudo-second-order model, including all processes of adsorption (external liquid film diffusion, surface adsorption, intraparticle diffusion and so on), could not accurately reflect the mechanism of this adsorption process 36 . For further exploring adsorption mechanism of TC on Ben/Rm/Ps-op, intraparticle diffusion model was employed to determine the type of rate-controlling step. This model can be delivered as follows: www.nature.com/scientificreports www.nature.com/scientificreports/ where k int (mg/(g•min 0.5 )) is the constants for the intraparticle diffusion model, and Const (mg/g) is a constant proportional to the extent of boundary layer thickness. Figure 9 expressed the linear plots involving two adsorption stages with different slopes. k int1 refers to the external adsorption rate constant in first step, and k int2 indicates the internal adsorption rate constant of the second stage by diffusion between particles into the adsorbent. The value of k int1 was higher than that of k int2 owing to a rapid increase in adsorption during the initial phase, with increased active sites available. This result is related

Model Parameter Ben/Rm/Ps-op ceramsite
Pseudo-first-order model www.nature.com/scientificreports www.nature.com/scientificreports/ to changes in mass transfer rate during adsorption process. The linear portion did not pass through the origin, suggesting that the adsorption mechanism of the TC onto Ben/Rm/Ps-op is not only restrained by the intraparticle diffusion step 37 .
Adsorption isotherm. Adsorption isotherm result at 20 °C was presented in Fig. 10. With the increase of the initial concentration of TC, adsorption capacity of Ben/Rm/Ps-op for TC increased overtly, while TC removal rate significantly descended. The isotherm results were further analyzed using linear, Langmuir, Freundlich, Tempkin and D-R (Dubinin-Radushkevich) isotherm models, as expressed below 38,39 : The linear isotherm model indicates that amount of adsorption is linearly proportional to the equilibrium solution concentration, which can be depicted as: The Langmuir isotherm model assumes that the adsorption sites on the surface of the monolayer are uniform and equivalent, with no interaction between adsorbate molecules at adjacent locations, which is expressed as 40 :   where n is the heterogeneity factor indicating the adsorption strength of the adsorbent, and K F (mg/g•(L/mg) 1/n ) is the constant in connection with the adsorption capacity. Tempkin and Pyzhev assumed that some indirect adsorbate/adsorbate interactions had effect on adsorption isotherms and suggested that the adsorption heat of all the molecules in the layer would decrease linearly with coverage due to these interactions. The Tempkin isotherm has been used as below: RT b A plot of q e versus lnC e could determine the constants A and B. The constant B is related to the adsorption heat 42 .
The D-R empirical equation put forward by Dubinin and Radushkevich, has been widely employed to depict the gases and vapours adsorption on microporous solids. In the case of liquid phase adsorption, several researches have indicated that the adsorption energy can be estimated via D-R equation. Assuming only monolayer adsorption occurs in micropores adsorption and the D-R equation is applicable, the adsorption capacity per unit surface area of the adsorbent at equilibrium, q e , can be described as 43 : where B is the constant related to the adsorption energy, q 0 is the ultimate capacity per unit area of adsorbent micropores, and ε is the Polenyi potential. The most probable adsorption energy, E, has been shown as: Fig. 11 depicted, the linear, Langmuir, Freundlich, Tempkin and D-R isotherm models were used to fit TC adsorption data onto Ben/Rm/Ps-op. The parameter values were based on the regression of the isotherm equations and were summarized in Table 7. The results showed that Langmuir isotherm model was not suitable for the adsorption, demonstrating that the adsorption of TC by Ben/Rm/Ps-op was not monolayer adsorption, neither Tempkin nor D-R isotherm models. Whereas, linear isotherm model was suitable for the determination of data due to the higher correlation coefficient R 2 (0.98689) than R 2 (0.95035) of Freundlich. For linear model, the coverage of monolayer and the initial amount of multilayer adsorption appear to be superimposed. Since there is no platform in linear model, it indicates that the adsorption does not reach saturation, and the multilayer adsorption thickness seems to increase indefinitely. www.nature.com/scientificreports www.nature.com/scientificreports/  www.nature.com/scientificreports www.nature.com/scientificreports/ Adsorption thermodynamics. Thermodynamic properties of TC onto Ben/Rm/Ps-op were further investigated, which could be described via Gibb's free energy (ΔG°), enthalpy (ΔH°) and entropy (ΔS°). The thermodynamic were estimated using the following relations 38,44 : c where R (8.314 J/K mol) is the gas constant; T (K) is temperature; K c is the equilibrium constant; C e is the equilibrium concentration of TC in the solution (mg/L); and C Ae is the amount of adsorbed TC on the adsorbent at equilibrium (mg/L). C Ae and C e are obtained from q e values of the pseudo-second-order model (Fig. 12). ΔS° and ΔH° were acquired from the slope and intercept of linear plot of lnK c versus 1/T according to Eq. (13) (Fig. 13). Figure 14 described adsorption capacity of ceramiste at different temperatures, and demonstrated that increasing temperature would promote adsorption capacity of TC by ceramsite. Table 8   www.nature.com/scientificreports www.nature.com/scientificreports/

Effects of pH, Ben/Rm/Ps-op dosage and ionic strength.
For the effect of pH on adsorption, the adsorption capacity and removal rate of TC increase within pH 2-7, and decrease during pH 7-10 ( Fig. 15a). Under strong acid condition (pH 2-4), TCH 3+ is the main form of TC, which could be combined with silicon dioxide with negative charge in ceramsite. While hydrogen ions in the solution are more competitive than TC in binding to the ceramisite voids at the same time. So the amount of TC adsorbed by ceramisite (from 1.86 to 1.94 mg/g) and removal rate (from 46.5% to 48.5%) did not increase significantly in the range of pH 2-4. With increasing pH from 4 to 7, the adsorption capacity of TC sharply increased by 0.62 mg/g and the removal rate climbed to 64% at pH 7. This because H + decreased in an order of magnitude, and the adsorption of TC by ceramsite increased obviously. However, from pH 7 to pH 10, the removal rate of TC by a large margin reduced to 53% with an adsorption capacity of 2.12 mg/g. During this period, the morphology of TC gradually changed from TCH 2° to TCH − , and the negative charge ratio on the surface of ceramsite increased, which were unfavorable to the adsorption of TC by ceramsite. This result indicates the adsorption capacity of Ben/Rm/Ps-op will be largely impacted by pH, and the adsorption performance was excellent at the neutral condition.
Besides, the adsorption of TC was conducted in the presence of different dosage of Ben/Rm/Ps-op. As Fig. 15b described, the removal rate of TC significantly increased from approximately 21.38% to 78.75% with increasing Ben/Rm/Ps-op dosage from 5 to 50 g/L. However, the TC adsorption capacity gradually decreased from approximately 3.42 to 1.26 mg/g. Given the efficiency and economy of such operation, the optimum Ben/Rm/Ps-op dosage is 20 g/L, under which both the adsorption efficiency and capacity were kept high.
In addition, adsorption experiments on the effect of ionic strength were conducted using 80 mg/L TC solution containing 0-0.25 mol/L NaCl or CaCl 2 at pH = 7 and the temperature of 20 °C. Figure 15c describes the adsorption behavior of TC versus ironic strength. The existence of NaCl (or CaCl 2 ) decreases the adsorption    www.nature.com/scientificreports www.nature.com/scientificreports/ capacity of TC onto Ben/Rm/Ps-op, which may be due to the competitive effect between Na + (or Ca 2+ ) and TC on the adsorption sites. Parolo et al. observed that it can be explained that metal cations in solution could easily chelate with TC 47 , and electrolyte can produce electrostatic shielding effect, thus affect adsorption 48,49 . In addition, increasing Na + (or Ca 2+ ) concentration can bring in the contraction of adsorbent pores, leading to that some adsorbate could not enter into pores, and the reduction of surface adsorption sites of Ben/Rm/Ps-op [50][51][52] . Further, it is clear that NaCl, a univalent electrolyte, had less negative impact on TC adsorption than a divalent CaCl 2 under identical conditions. Thus, it can be concluded that coexisting ions had adverse effect for TC adsorption onto Ben/Rm/Ps-op.
Dynamic adsorption of TC. The effect of hydraulic retention time (HRT) and packing height on TC removal were investigated, and the result was shown in Fig. 16. It can be seen from the figure that HRT had a great influence on the removal of TC by Ben/Rm/Ps-op. When HRT = 5, 10 and 15 h, the average removal rates of TC by ceramsite got to 69.0%, 77.7% and 81.1% respectively. With the increase of HRT, the removal rate of TC by ceramsite increases. The reason is that the increase of retention time of solution through the packed column will lead to more sufficient contact reaction between ceramsite and TC, which makes the total amount of TC adsorbed by ceramsite increase. In addition, with the increase of HRT, the amount of TC adsorbed by ceramsite increases, but the degree of increase decreases (77.7-69.0% > 81.1-77.7%). This could be interpreted as that with the prolongation of adsorption time, the adsorption sites decrease and the adsorption difficulty increase. Since the HRT of a CWs system is usually longer than 3 days 53 and the removal efficiency have already reached 81.1% at HRT = 15 h, the Ben/Rm/Ps-op has the high potential to effectively remove TC as the CWs substrate under a dynamic flow condition. www.nature.com/scientificreports www.nature.com/scientificreports/ Figure 16b presents TC concentration of Ben/Rm/Ps-op packed column at different packing heights. The TC concentration at different packing heights showed a similar change trend versus the operation time, i.e., rapidly increasing in the initial days, then reaching a relatively stable level, and gradually increasing in the late stage. However, the initial rapidly increasing takes different time. TC concentration quickly increase in the initial 5, 7 and 8 days, respectively for height = 10, 15, 20 cm. This may be attributed to that with the increase of the quality of adsorbent, the adsorption sites on the surface of adsorbent increase, which lead to prolonging the adsorption time of TC. We could also see in Fig. 16b, the lower section played the key role in the adsorption of TC, and it also arrive at saturation first during the continuously operation. Fig. 17, good adsorption capability of the regenerated adsorbent was still retained after three rounds of sorption-desorption cycles. In addition, TC removal rates for repeated three times were 61.7%, 58.4% and 53.2% respectively. The decrease (2.3-5.2% in every cycle) of removal rate might have been owing to the loss of irreversible occupation of partial-adsorption sites 54 . Nevertheless, the adsorption amount of ceramsites still remained at a high value (2.13 mg/g, C 0 = 80 mg/L) after three consecutive cycles, suggesting the high reusability capability and stability of Ben/Rm/Ps-op for TC removal.  www.nature.com/scientificreports www.nature.com/scientificreports/ conclusions In this study, a kind of CWs Ben-Rm-Ps ceramsite was prepared to remove TC in effluent. Ben/Rm/Ps-op ceramsite was prepared with the condition Ben: Rm: Ps = 4:1:0.9, preheating temperature = 240 °C, preheating time = 20 min, calcining temperature = 1150 °C, and calcining time = 14 min, which possessed microporous structure and low heavy metal leaching toxicity. The second-order kinetics and linear isothermal model can well simulate the adsorption of TC by Ben/Rm/Ps-op, and the maximum adsorption capacity can reach for 2.5602 mg/g. In addition, TC adsorption onto Ben/Rm/Ps-op was demonstrated a spontaneous endothermic process and higher temperature enhanced the adsorption. Further, Ben/Rm/Ps-op has been also proved the high potential to effectively remove TC as the CWs substrate under a dynamic flow condition and high reusability capability and stability for TC removal under batch tests.

Regeneration of Ben/Rm/Ps-op. Seen in
This study combines basic theory and engineering application research, and has important value for the research and development of CWs matrix filler. The research results of this subject will provide an important reference for the research and application of artificial calcined ceramsite as a light aggregate in water pollution control. In addition, the preparation and application of new ceramsite matrix can not only enhance the pollutant removal function of CWs, but also utilize solid wastes such as red mud and biomass. Therefore, the research and development of ceramsite products also has comprehensive economic, social and environmental benefits.