1 Introduction

One alarming result of the escalating growth in the global population is the pollution of the surface and underground water, which has placed a ceiling on the demand for fresh water, with repercussions on the environment and health [1, 2]. Antibiotics rank high as the commonest and most hazardous pollutants present in pharmaceutical wastewater [3]. Reports from several research studies, over the last decade, show that the quantity of antibiotics produced in pharmaceutical factories indicate an annual rise from 100 thousand tons to 200 thousand tons [4,5,6]. Of grave concern is that these antibiotics are crucial compounds, extensively used to treat many bacterial diseases occurring in both humans and animals. When antibiotics are carelessly handled or overused, they can, in due course, be found present in the water systems. When these enter the drinking water sources, as well as food crops, they will cause disastrous consequences to public health, and raise the resistance bacteria to antibiotics, now acknowledged as an environmental problem across the world. Besides, when the antibiotics permeate into the soil and ground water, they have detrimental effects on agricultural crops. This affects the plant growth in terms of the ability to impede respiration and impact microbial growth, as well as interfere in the process of nitrification and iron recovery (III). Therefore, there is urgent need to treat and minimize the negative influence exerted by the antibiotics present in the environment [7], particularly in antibiotic-rich wastewater [7,8,9].

An antibiotic of the Fluoroquinolone class, Ciprofloxacin (CIP: C17H18FN3O3), has broad-spectrum antibacterial action, and is extensively employed in the treatment of aerobic bacterial infections. Therefore, it is very important that effective processes are available to remove the CIP from the aquatic environment and nullify its inherent negative influences [10,11,12].

Antibiotic-removal from wastewater has been accomplished using a variety of biological treatments like chlorination, filtration, coagulation and sedimentation, photo-transformation, ozonation, or processes of advanced oxidation [13,14,15,16,17]. Despite their effectiveness in antibiotic-removal, these methods are not very popular, as they involve high cost, complexity, or release by-products [18, 19].

The adsorption method has been identified as superior to the others cited because of its simple application, effective pollutant removal, and more rapid action; above all, it is easy to use in field conditions and in treatment systems [20,21,22]. The literature search revealed several works investigating the pharmaceutical-adsorption on activated carbon (AC), clay, carbon nanotubes, silica, and zeolite [23, 24]. The adsorbent with the greatest efficacy is AC, which adsorbs the pharmaceuticals very well, as it has a highly porous and large surface area with the capacity of adsorbing a variety of organic and inorganic elements, even under an extensive range of conditions [25, 26]. Despite the effectiveness of AC and other mesoporous carbon materials as successful adsorbents in CIP-contaminated water, the issue of pollution relentlessly persists because it is hard to actually separate them from the treated wastewater [27, 28]. In reality, the removal of the used adsorbent poses one of the most critical factors from an economical and operational standpoint in the treatment process, particularly in the adsorption method. To find a solution for the separation issue, as well as raise the efficiency of removal through the increase in surface area, one promising method appears to be magnetization of the used adsorbent [29, 30]. Recently, magnetizing the adsorbents using Fe3O4 magnetic nanoparticles as a coating was extensively used to heighten the degree of adsorption of the AC and other similar adsorbents, thus enhancing pollutant removal [21, 31, 32].

Although the AC has been identified as the adsorbent with the highest efficiency employed in antibiotic removal, it is not cost-effective on a large scale with regards to synthesis and regeneration. After much research over the last decade, more inexpensive adsorbents have been found to be very effective, which include a variety of non-traditional adsorbents such as rice husk ash, eucalyptus bark, sawdust, Lemna minor (LM), canola, and Azolla filiculoides [33, 34]. In fact, LM, a freshwater plant, the common duckweed, belonging to the genus Lemna, has been identified. As these leaves grow, the plants themselves divide to stand as separate individuals [35, 36]. They grow in freshwater ponds and gently-flowing streams. The LM, also effective as animal feed, is good in nutrient recovery from wastewater, besides its other applications. In the present study, this was the preferred choice, as it could be used a very cost-effective adsorbent [37, 38].

Therefore, the aim of the current study was to synthesize Fe3O4-ACLM and investigate its efficiency to adsorb the CIP molecules from aqueous solutions. The characterization parameters relevant to the adsorption process were determined using advanced techniques. Furthermore, the isotherm, kinetic, and thermodynamic processes of CIP adsorption onto Fe3O4-ACLM were discussed.

Therefore, the objective of the study presented here was to synthesize Fe3O4-ACLM and explore its efficacy in the adsorption of the CIP molecules from aqueous solutions. The characteristic parameters concerned with the adsorption process were identified with the use of advanced techniques. Furthermore, a discussion is presented on the isothermal, kinetic, and thermodynamic processes of CIP adsorption onto the Fe3O4-ACLM. Employing five statistical parameters, different isotherms and kinetic models were used to fit the experimental data. Testing was done on the recyclability of the adsorbent in five consecutive adsorption–desorption processes of the CIP.

2 Material and methods

2.1 Chemicals

The Sigma–Aldrich company (Germany) provided the CIP with molecular weight and purity of 385.82 g/mol and 99%, respectively; as well as Acetonitrile of 99.8%, and phosphoric acid of 99.9% purity. Merck (Germany) supplied the other chemicals such as acetic acid (≥ 99.7%), FeCl3.6H2O (≥ 98%), FeCl2.4H2O (≥ 99%), HCl (36.5%), and NaOH (≥ 98%).

2.2 Preparation and characterization of the Fe3O4-ACLM adsorbent

The LM, sourced from the Anzali Wetland, Iran, was used in the preparation of the activated carbon (AC). First, the LM was rinsed using pure water and dried for 1 week, at room temperature. Next, it was subjected to milling and manual sieving through physically shaking the stainless-steel mesh plates having the standard mesh size of 0.45 mm. Therefore, for the AC production, the LM was dried and crushed and 5 g of this was taken (after checking the weight on a digital scale). It was then stirred for 6 h in 40 ml of 28% ZnCl2 solution. Next, it was placed in an electric oven at 105 °C for 24 h and dried. This mixture was then placed in a vertical stainless-steel reactor though which, high purity N2 gas at 300 cm3/min flow rate, was passed. The temperature was adjusted to 450 °C and activation was done for 1 h at this temperature.

The excess ZnCl2 was eliminated from the LMAC thus produced, utilizing HCl (0.5 M). The next steps involved filtration and washing with hot water, and final drying (at 105 °C for 24 h). At first, 4 g of LMAC was added to a flask, which was kept for 3 h in an ultrasound bath at 80 °C, to which 20 ml HNO3 solution (1 M) was poured in. Three hours later, a separation of the LMAC adsorbent from the solution was observed. It was filtered through a filter and dried for 1 day, at room temperature. The Fe3O4 nanoparticles and the LMAC, in the ratio of 3:2, were then coagulated in 200 ml distilled water. This was transferred to an ultrasound bath for 1 h maintaining the temperature at 80 °C. The Fe3O4-ACLM thus prepared was then filtered and thrice washed with distilled water and rinsed well once with ethanol (200 ml). Finally, the Fe3O4-ACLM was oven-dried for 24 h at 100 °C.

The excess ZnCl2 was removed from the LMAC produced, with the aid of HCl (0.5 M). Next, the LMAC was filtered and washed using hot water and ultimately dried for 24 h (at 105 °C). Now, 4 g of the LMAC was put in a flask and kept in an ultrasound bath for 3 h at 80 °C, to which 20 ml of HNO3 solution (1 M) was added. Three hours later, separation of the LMAC adsorbent from the solution was done through filtration, and dried for 1 day at room temperature. Next, coagulation of the LMAC was done using the Fe3O4 nanoparticles in a 2:3 ratio in 200 ml of distilled water. This was then kept in an ultrasound bath for 1 h at 80 °C. The Fe3O4-ACLM thus prepared was filtered and thrice washed using distilled water and rinsed once using 200 ml of ethanol. Finally, the Fe3O4-ACLM was oven-dried for 24 h maintaining the temperature at 100 °C.

The TEM model (LEO 912 AB) was used in this study, at 100 kV, and the SEM/EDX (MIRA3 FEG-SEM-Tescan) was also employed to ascertain the size, shape, and morphology of the Fe3O4-ACLM. Then, using the KBr pellet method, the FTIR (Thermo Nicollet AVATAR5700) was also employed, keeping the range between 4000 and 400 cm−1. Using the nitrogen adsorption at – 196 °C in a Micromeritic Model ASAP 2020 Plus, the specific surface area and pore size values were evaluated through the Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) analyses. Further, with the help of a vibrating sample magnetometer (VSM) (Micromeritics Instrument Corp., Norcross, GA, USA), the magnetization curves were identified at 300 K, in a 50 KOe applied field.

2.3 Batch adsorption tests

The batch mode was used for the CIP adsorption studies. The adsorption experiments were conducted using 250 ml Erlenmeyer flasks containing different concentrations (25–100 mg/L) of the CIP ionic solution. First, 0.5 g of pure CIP powder was dissolved in 1 L of distilled water to produce 500 mg/L of CIP stock solution. The stock solution was diluted to prepare the test concentrations required. Next, the necessary quantity of Fe3O4-ACLM was added to the solution. The flasks were placed on an electrical thermostatic shaker at 120 rpm and stirred. The effect exerted by the various experimental parameters for CIP removal were examined, keeping the pH in the 3–11 range. The parameters included the dose of Fe3O4-ACLM (0.1 to 1 g/L), initial CIP concentration (25–100 mg/L), contact time (10–150 min), and temperature (20–50 °C).

The experiments were performed adopting the optimization method, so that when only one parameter was changed, while keeping the others constant, the optimal value of the variable parameter was easily calculated. In this study, both the initial concentration and contact time were optimized and tested using different concentrations, from 25 to 100 mg/L, as well as different contact times (10–150 min), but taking care to maintain the adsorbent dose, pH, and temperature at constant. Once the optimal concentration and contact time were obtained, the other parameters were examined at that optimal concentration and contact time.

After the Fe3O4-ACLM was separated using a strong magnet, the supernatant was drawn off with the help of a 0.22 µm polypropylene membrane syringe filter. All the tests were done twice, using the mean data to record the results. The CIP in all the samples was ascertained using the column HPLC (C18 ODS) and a 2006 UV detector at 277 nm. The mobile phase included 0.05 M phosphoric acid/acetonitrile with 87/13 volume ratio and 1 ml/min flow rate of the injection. The solution pH was considered adjusted by using either 0.1 M HCl or 0.1 M NaOH.

Calculation of the adsorption capacity (qe) and removal efficiency (% Removal) using general Eqs. 1 and 2 [39]:

$${q}_{e}=(\frac{{C}_{0}-{C}_{e}}{\mathrm{M}})\times \mathrm{v}$$
(1)
$$\mathrm{\% Removal}=(\frac{{C}_{0}-{C}_{e}}{{C}_{0}})\times 100$$
(2)

where qe refers to the mg of the absorbed CIP on 1 g of Fe3O4-ACLM; V (L) indicates the volume of the solution which contains the CIP, in contact with the Fe3O4-ACLM; Co and Ce imply the initial concentration and equilibrium (residual) of CIP in the solution (mg/L), respectively; and M (g) is the quantity of the added Fe3O4-ACLMinto the solution.

2.4 Error analysis

The model most compatible with the experimental data was determined by ascertaining and comparing the correlation coefficients (R2). Thus, to identify the most suitable model, the R2 value is the best criterion; however, there are limitations in solving the nonlinear isotherm models because they require the utilization of different parameters. Hence, in this study, in order to determine the most appropriate model, five different error functions were employed. The error functions, used in this study, therefore are expressed in the equations given [40,41,42].

$$\begin{array}{cc}\mathrm{SSE}=\sum_{\mathrm{i}=1}^{\mathrm{n}}({q}_{\mathrm{e cal}}-{q}_{\mathrm{e meas}}{)}_{\mathrm{i}}^{2}& \mathrm{Sum of squared errors}\end{array}$$
(3)
$$\begin{array}{cc}\mathrm{SAE}=\sum_{\mathrm{i}=1}^{\mathrm{n}}({q}_{\mathrm{e cal}}-{q}_{\mathrm{e meas}}{)}_{\mathrm{i}}& \mathrm{Sum of absolute errors}\end{array}$$
(4)
$$\begin{array}{cc}\mathrm{ARE}= \frac{100}{\mathrm{n}}\sum_{\mathrm{i}=1}^{\mathrm{n}}(\frac{{q}_{\mathrm{e cal}}-{q}_{\mathrm{e meas}}}{{q}_{\mathrm{e meas}}}{)}_{\mathrm{i}}& \mathrm{Average relative error}\end{array}$$
(5)
$$\begin{array}{cc}\mathrm{HYBRID}=\frac{100}{\mathrm{n}-\mathrm{p}}\sum_{\mathrm{i}=1}^{\mathrm{n}}\left[\frac{{(q}_{\mathrm{e meas}}-{q}_{\mathrm{e cal}{)}_{\mathrm{i}}^{2}}}{{q}_{\mathrm{e meas}}}\right]& \mathrm{Hybrid fractional error function}\end{array}$$
(6)
$$\begin{array}{cc}\mathrm{MPSD}=100\sqrt{\frac{1}{\mathrm{n}-\mathrm{p}}\sum_{\mathrm{i}=1}^{\mathrm{n}}(\frac{{q}_{\mathrm{e meas}}-{q}_{\mathrm{e cal}}}{{q}_{\mathrm{e meas}}}{)}_{\mathrm{i}}^{2}}& \mathrm{Marquarts percentage standard deviation}\end{array}$$
(7)

where \({q}_{\mathrm{e meas}}\) and \({q}_{\mathrm{e cal}}\) are the experimental and calculated uptakes, respectively.

3 Results and discussion

3.1 FTIR analysis

FTIR was one of the analyses performed for the adsorbent synthesized in this study, i.e., the Fe3O4-ACLM, performed in the 4000–400 cm−1 range (Fig. 1a). The dried sample was first placed on a silicon substrate, transparent to the infrared, and the spectra were measured according to the transmittance method [43]. From the FTIR spectrum, evidence of a very strong band was visible at 3312 cm−1; this is attributed to the O–H stretching of the hydrogen-bonded hydroxyl groups. The absorption peak at 3472 cm−1 may be caused due to the N–H stretching which implies the presence of free and intermolecular-bonded hydroxyl groups resulting from the alcoholic or phenolic functions and amine group. The findings of this analysis indicate the likelihood of an absorption band at 1704 cm−1; this suggests the stretching vibration caused by the carboxyl group (C = O), related to the acid molecule, which becomes adsorbed onto the surfaces of the composites. At 1619 cm−1, the peak noticed was found to be linked to the carboxylate (COO-) stretching vibration which could be due to the lignin aromatic group; a band present at around 1474 cm−1 is supposed to be due to the –CH2 deformation bending. The strong -C-O band visible at 1041 cm−1 resulting from the –OCH3 group is because of the lignin structure present in the LM head, the characteristic peak of the polysaccharides [32]. The peaks observed at 894 cm−1 and 571 cm−1 may be caused due to the vibrations of the Fe–O bond for FeO(OH), as well as the Fe–O bond, respectively, which lends support to the Fe3O4-ACLM nanoparticles present there.

Fig. 1
figure 1

The FTIR (a), TEM (b), SEM (c, d), BET (e), the magnetic hysteresis loops (f), and XRD (g) analysis

Full size image

3.2 TEM images

The TEM images of Fe3O4-ACLM are shown in Fig. 1b. The TEM images of the Fe3O4-ACLM (1b) thus prepared indicate the very porous structure of adsorbent, with an abundance of both micropores and mesopores. Further, the Fe3O4 NPs are extensively dispersed on the ACLM surface and reveal an almost spherical morphology, with diameter of almost uniform size of 9.8 ± 1.5 nm, in fact concurring with the BJH analysis.

3.3 SEM images

The SEM images shown in Fig. 1c, d reveal the morphology and microstructure of the Fe3O4-ACLM prior to and post the adsorption process; according to this, several pores and holes are present on the external Fe3O4-ACLM surface, suggesting the existence of a variety of sizes and shapes for the magnetite Fe3O4-ACLM. Further, prior to the CIP adsorption, the Fe3O4-ACLM appeared to have irregular shape and a rough surface, as shown in Fig. 1c. However, post the adsorption, the Fe3O4-ACLM showed a flatter surface morphology, as shown in Fig. 1d, caused by the CIP covering the Fe3O4-ACLM adsorbent surface. Besides, a few white particles evident on these sample surfaces were most likely iron oxide nanoparticles, implying the creation of a layer of neatly dispersed iron particles on the surface of the adsorbent. Identical features were also reported by Balarak on the activated carbon surface derived from Azolla filiculoides [34].

3.4 BET analysis

From the BET analysis, it was evident that at 582 m2/g the specific surface area of the Fe3O4-ACLM was a little smaller than that of the ACLM (645 m2/g). The reason for this condition is due to the Fe3O4 covering the AC surface; in fact, this resulted in the slightly lower specific surface area of the Fe3O4-ACLM adsorbent. The BET analysis (Fig. 1e) was done to investigate the microstructure of the Fe3O4-ACLM. According to the IUPAC classification, a typical type IV isotherm was obtained for the Fe3O4-ACLM, which is characteristic for porous materials. From the study of the BJH pore size distribution, the recovered Fe3O4-ACLM showed the presence of pores, having diameters in the 8 to 12 nm range.

A summary of the results regarding the properties of the Fe3O4-ACLM and ACLM are shown in Table 1, which include moisture content, total pore volume, and a number of various elements, etc.

Table 1 Characteristic of the Fe3O4-ACLM and ACLM
Full size table

The magnetic hysteresis loops of the Fe3O4-ACLM and Fe3O4 are revealed in Fig. 1f. From these results, it was clear that the value of the magnetization saturation (Ms) for the Fe3O4-ACLM nanoparticles was 37.6 emu/g, whereas for the Fe3O4 nanoparticles it was 81.2 emu/g. The characteristic superparamagnetic behavior was detected for both materials, recognized by the typically shaped curve [38]. On comparing the results found for the Fe3O4-ACLM and Fe3O4, a reduction in the magnetic response for Fe3O4-ACLM was noted, most likely due to the layers of magnetic material present in the carbon structure. Hence, the Fe3O4-ACLM and Fe3O4 (having high Ms values) can reveal a quick response to the external magnetic field, which in turn will cause these adsorbents to rapidly separate from the aqueous CIP solution.

3.5 XRD analysis

The X-ray powder diffractograms of the Fe3O4-ACLM and Fe3O4 nanoparticles are seen in Fig. 1g. These patterns reveal a series of typical peaks (220), (311), (400), (111), (511), and (442), which concur well with the inverse cubic spinel phase of the Fe3O4 (magnetite, JCPDS card no. 85–1436). Using Scherrer’s formula, the mean diameter of the crystallite deduced from the diffractogram by [40] is 11 nm. This confirms the size perceived by the BJH distribution of pore size. Additionally, from the XRD pattern, the characteristic peaks of carbon, of about 27°, were observed at 2θ; this supports the fact that amorphous carbon is present. From the intensity of the peaks and their rather narrow widths, a high level of crystallinity is indicated in the nanocomposite [39].

3.6 Effect of different parameters on CIP adsorption

The influence exerted by the initial CIP concentration and contact time are evident in Fig. 2a. As cited, any increase in the concentration of the initial CIP is related to an improvement in the capacity of the CIP to be adsorbed onto the Fe3O4-ACLM. At the initial contact time, rapid adsorption was noted, connected to the abundance of the available active sites on the surface of the adsorbent material [40]. As these active sites were gradually occupied, efficient adsorption was observed to decrease. Further, the increase in the initial CIP concentration is linked to the rise in the quantity of the CIP adsorbed; this occurs because of the availability of the higher gradient of available molecules in the medium, which caused a rise in diffusion during the process of mass transfer [41]. Although the data collection was done at the contact time of 150 min, the equilibrium time of the adsorption process took only 75 min. The findings from this study concurred with those reported in earlier studies [44, 45].

Fig. 2
figure 2

The effect of the initial CIP concentration (a), adsorbent dose (b), pH (c), and pHpzc analysis (d)

Full size image

In Fig. 2b, it is evident that the removal percentage (%) and adsorption capacity (qe) of CIP is given against the dose of the Fe3O4-ACLM. The results implied an increase in the removal percentage in response to an increase in the Fe3O4-ACLM dose; this occurs due to the greater number of active sites available for adsorption. However, when the quantity of the Fe3O4-ACLM is increased, the adsorption capacity (qe) is observed to decrease; a reduced quantity of the CIP adsorbed per gram of Fe3O4-ACLM could be regarded as the cause for this occurrence. Another plausible reason could be the adsorption sites overlapping or aggregating, which ultimately resulted in reducing the total specific surface area of the Fe3O4-ACLM available and promoting the route for diffusion [41].

The influence exerted by pH on the CIP adsorption by Fe3O4-ACLM, as well as its optimum, was determined to be in the 3.0–11.0 range, and the results are shown in Fig. 2c. In the course of conducting this part of the experiments, the constant values of the initial concentration of the CIP (25 mg/L), Fe3O4-ACLM dose (0.6 g/L), temperature (30 ± 2 °C), and contact time (75 min) were taken into account. The adsorbent in this study revealed a pHpzc value of 6.4 (Fig. 2d). From these results, it was evident that the Fe3O4-ACLMto anionic molecules tended to be greater when the pH value dropped below the pHpzc, a phenomenon understood as an increase in the cations on the carbon surface. In fact, when the pH values exceed that of the pHpzc, there is heightened adsorption of the cationic compounds [30]. When the CIP gets disassociated in water, the anionic molecules are formed; it is, therefore, anticipated that, at pH values less than the pHpzc, the Fe3O4-ACLM acts as a potential CIP adsorbent. On comparison of the removal efficiencies at different pH values, the CIP adsorbed starts to reduce as the pH rises. For the amine and carboxylic acid groups in the piperazine moiety, the pKa values for the CIP are 8.7 and 6.1, respectively. When the pH of the solution drops below 6.1, the CIP is observed most often as a cation; however, when the pH hovers in the range of 6.1 to 8.7, the CIP is zwitterionic. When the pH of the solution exceeds 8.7, the CIP loses the protons of the carboxylic group, and appears mostly as anions. When the pH plummets to less than 6.1, there is a rise in the CIP removal because the electrostatic attraction increases due to the opposite charges between the CIP and Fe3O4-ACLM. Earlier studies done on the CIP adsorption employing different adsorbents reported similar findings [7, 9].

3.7 Isotherm studies

The adsorption equilibrium isotherm is usually used to explain the adsorption properties. In the present study, the Langmuir, Freundlich, Temkin, and D-R isotherms were used to investigate the equilibrium adsorption of the CIP by Fe3O4-ACLM (Table 2). To gain greater knowledge regarding the theoretical background of the isotherm model, the following references have been referred [25, 46]. As evident from Table 2, the maximum adsorption is achieved with the rise in temperature, with better performances revealed at the high temperatures, also concurrent with the thermodynamic adsorption findings. From these isotherm studies, the R2 values obtained for a model under investigation for the present adsorption system is in the order as given: Langmuir > D-R > Temkin > Freundlich. According to the order mentioned, for the present study, the higher R2 values (0.995–0.998) were linked to the Langmuir isotherm model, suggesting the incidence of the monolayer coverage of the pollutant molecules on the surface of the Fe3O4-ACLM. At different temperatures, the qm values obtained for the Langmuir isotherm equation were 134.2, 149.5, 161.4, and 178.7 mg/g, respectively.

Table 2 Isotherm parameters for adsorption of CIP on Fe3O4-ACLM at various temperatures
Full size table

For the Temkin model, the values of B were found to be below 1 signifying that the adsorption process of the CIP onto the Fe3O4-ACLM is endothermic in nature, at the concentration under investigation. Finally, based on the mean adsorption energy (E) drawn from the D-R isotherm model, it was evident that there was a significant part for the hydrophobic effect, hydrogen bonding, electrostatic attractions, and interaction between Fe3O4-ACLM and CIP to play, in this adsorption process.

Also, for the D-R isotherm, parameter E dropped below 8 kJ/mol at the low temperatures and exceeded 8 kJ/mol at the high temperatures, suggesting that at low temperatures the CIP adsorption is a physical process, while at the higher temperatures the chemical reactions are the ones assisting in the adsorption process [35]. Also, for the Langmuir isotherm at each temperature, the RL parameter was found to hover between zero and one, indicating that the uptake of the CIP by the Fe3O4-ACLM was optimal [47].

In Table 2, the results of the error analyses for the different isotherms are shown and as is evident, the Langmuir isotherm, at all temperatures, reveals a higher regression coefficient and a lower error coefficient. In Table 3, a comparison is made of the maximum adsorption capacity as calculated from the Langmuir isotherm for the CIP adsorption with a similar research study done on the elimination of this antibiotic; of note, a large percentage of these studies have been conducted over the recent years, and the findings reveal Fe3O4-ACLM to have the highest efficiency when compared to other similar adsorbents, as displayed in Table 3.

Table 3 Comparison of the maximum adsorption values of various adsorbents for CIP
Full size table

3.8 Adsorption kinetics

From Fig. 3a, it is evident that the initial adsorption rate was so high that 80% of the adsorption occurred at 25 mg/L concentration within the first 30 min of the process; this was caused by the high concentration gradient and the absence of adsorption by the adsorbent at the commencement of the process [44]. Moderate processing speed was seen from 30 to 60 min, which escalated in a gentle gradient, and no adsorption was noted from 60 min onwards, the reason for the reduction at the adsorption sites.

Fig. 3
figure 3

The effect of the initial CIP concentration on the removal % a pseudo-first-order kinetic model, b pseudo-second-order kinetic model, c intraparticle diffusion model, and d of the CIP adsorption on Fe3O4-ACLM

Full size image

Using adsorption kinetics, the reaction rate and paths can be understood, and are dependent upon the physical and chemical characteristics of the adsorbent. Three models were employed to explain the adsorption kinetics, namely the pseudo-first-order kinetics (PFO), pseudo-second-order kinetics (PSO), and intra-particle diffusion (IPD) models [55, 56].

In fact, Table 4 reveals the equations for the kinetics, in addition to the analysis of the results. As shown in Table 4 and Fig. 3b, c, the equilibrium data at all the concentrations investigated appear to satisfy the PSO model more than the PFO kinetic model. Further, the error coefficients for the PSO model were lower than those of the PFO kinetic one, and finally, the results recorded from calculating the experimental qe and qe drawn from the PSO model show closeness to each other, revealing smaller number of differences and greater consistency. Notably, the results of the qe exp for the concentrations of 25, 50, 75, and 100 mg/L were found equal to 41.6, 78.25, 105.6, and 127.1 mg/g, respectively; however, a comparison of these numbers with qe assessed from the PFO model, as given in Table 2, is very different. Therefore, it is evident that in all respects the equilibrium data follow the PSO kinetics.

Table 4 Kinetic parameters for the CIP adsorption on Fe3O4-ACLM
Full size table

The IPD was used to study the adsorption mechanism and the equation is given in Table 4. The degree of effectiveness of the boundary layer thickness is determined from the value of C in the IPD equation. From this, it appears the higher C values indicate the important impact exerted by the boundary layer. When the adsorption process is under the direction of the IPD, plotting the qt versus t1/2 (Fig. 4d) results in a linear plot.

Fig. 4
figure 4

Effect of temperature on qe (a), ionic strength effect on CIP adsorption on Fe3O4-ACLM (b), and Fe3O4-ACLM recyclability experiment (c)

Full size image

The IPD is introduced as a rate-controlling step only if the lines obtained cut through the origin. Else, other factors, like the boundary layer diffusion (or external mass transfer), barring the IPD, exert influence on the adsorption process. From Fig. 3d, it is evident that there are three different phases of the process, namely surface sorption, intra-particle diffusion, and a likely chemical reaction step. The last of the stages cited above is insignificant and occurs rapidly. Table 4 lists data including the k and C values for all the three stages, obtained by fitting the data acquired to the IPD model. A critical factor that has been introduced is the driving force for the adsorption processes. The boundary layer diffusion also is crucial to the adsorption process as the regression of qt versus t1/2 has not cut through the origin. The higher C values acquired for all the kinetic studies indicate the role of the IPD in the process of adsorption. However, barring the IDP, there are other factors which also control the process under investigation [57, 58].

3.9 Effect of temperature and determination of thermodynamic parameters

From the findings arrived at in this study, the highest removal percentage was observed at 323 K, indicating that the adsorption process was endothermic (Fig. 4a). When the ions increased in mobility in response to a rise in the temperature, the interaction between the ions and the Fe3O4-ACLM surface was enhanced, which may be the likely cause for the results in this section [59].

Through the use of thermodynamic studies, the physical or chemical nature of the adsorption mechanism is understood. In order to calculate the elimination of CIP by Fe3O4-ACLM, the thermodynamic parameters, namely standard Gibb’s free energy (ΔG0), standard enthalpy (ΔH0), and standard entropy (ΔS0), were used, as shown in the equations given here [60].

$${\Delta \mathrm{G}}^{0}=-\mathrm{RT Ln K}$$
(8)
$$\mathrm{Ln K}=\left({\Delta \mathrm{S}}^{0}/\mathrm{R}\right)-\left({\Delta \mathrm{H}}^{0}/\mathrm{RT}\right)$$
(9)

where R (8.314 J/mol K) is the universal gas constant; K refers to the constant drawn from the Langmuir equation, and T implies the absolute temperature (K). The thermodynamic parameters for the CIP adsorption onto the Fe3O4-ACLM surface at various temperatures are noted in Table 5.

Table 5 Values of thermodynamic parameters for the adsorption of CIP on Fe3O4-ACLM
Full size table

The possibility of the adsorption process and its spontaneous nature are indicated by the negative value of ΔG0 in Table 5. This concurs with the results attained from the Langmuir RL separation factor, where the RL values hover from 0 to 1 (0 < RL < 1) and the Freundlichn symbol exceeds 1. Notably as the temperature increased, the value of ΔG0 reduced, implying that the adsorption process is enhanced by a rise in temperature [61]. When the ΔH0 value is positive, it indicates the endothermic and irreversible nature of the adsorption process [62]. Further, the positive values of ΔS0 also suggest the affinity of the Fe3O4-ACLM surface for the CIP ions and the high degrees of disorder and randomness at the solid-solution interface [63].

3.10 Ionic strength test

From prior investigations, ion strength apparently affects the electrostatic interactions between the adsorbents and pollutants. In the meanwhile, electrolytes also influence the adsorption process as they compete with the pollutant ions for adsorption onto the surface of the adsorbent [58]. The effect exerted by this parameter on the CIP adsorption onto the Fe3O4-ACLM is evaluated, as shown in Fig. 4b which suggests that ionic strength has a negative relationship with the capacity for CIP adsorption.

The reason for this, most likely, is that when the pH is acidic, any increase in the concentration of the other ions raises the competition for adsorption onto the adsorbent surface, while simultaneously neutralizing the positive charge carried by the adsorbent surface; in addition, there is a rise in the electrostatic repulsion between CIP molecules to be adsorbed and the surface of the Fe3O4-ACLM adsorbent [59].

3.11 Reusability of Fe3O4-ACLM

Recovery and reusability are the critical parameters assessed for the selection of an adsorbent which is cost-effective and practical for use in pilot-scale treatment systems. The reusability of the adsorbent synthesized in the present study was considered, i.e., Fe3O4-ACLM was assessed by performing CIP adsorption cycles (CIP = 25 mg/L). From Fig. 4c, the results drawn from the experiments done to evaluate the recyclability and stability of the Fe3O4-ACLM adsorbent can be seen; all the experiments were performed within 75 min of contact time. Recycling and reuse of the adsorbent synthesized in this work can be performed easily using a strong magnet, post treatment. In our findings, a drop from ~ 100 to ~ 91% (by only 9%) in the adsorption activity of Fe3O4-ACLM was identified after six successive cycles. Two reasons have been put forward for this marginal reduction in the adsorption activity. In the recovery phase, when washing and drying are done, material losses may take place, resulting in a decrease in the adsorbent dose in the next cycle, which causes the surface adsorbent activity to decline, thus lowering the efficiency. Further, the likely alterations in aggregation and fouling properties of the magnetic nanoparticles in those six cycles can be understood as the other reason for the event cited above. Based on the documents available, aggregation can cause the effective surface area to decrease and the number of active sites to reduce.

4 Conclusion

The present study included the preparation of activated carbon from the Lemna minor plant which was magnetized using Fe3O4 nanoparticles, done using both chemical and physical techniques. The adsorbent thus prepared was investigated for its efficacy in the elimination of the antibiotic ciprofloxacin (CIP) from aqueous solutions. Here, ZnCl2 was used as the dehydration agent in the chemical activation process; the high temperature heating process was employed for the physical activation. The maximum CIP removal efficiency (100%) was achieved under the following conditions: initial CIP concentration = 25 mg/L, solution pH = 3, Fe3O4-ACLM dose = 0.75 g/L, and adsorption time = 75 min. From the kinetic results, the involvement of more than one mechanism was evident in the adsorption process, with the possibility of intra-particle diffusion and film diffusion control also playing a role in the adsorption process. Besides, it was found that the pseudo-second order models showed higher suitability to describe the kinetics data, a fact confirmed by the findings of the parameters of goodness-of-fit. From the experiments of sequential adsorption and desorption, it was clear that the successful reusability of the Fe3O4-ACLM was possible for six cycles, with the adsorption efficiency revealing only 9% reduction. From the linear regression with five error functions showed it was clear that the Langmuir model could be used to describe the CIP adsorption on Fe3O4-ACLM. The thermodynamic investigation revealed the endothermic aspect and spontaneity of the adsorption process. Using the Langmuir model, the maximum adsorption capacity of the Fe3O4-ACLM for the removal of the CIP molecules was 178.8 mg/g, a higher value than for a few adsorbents employed earlier in CIP elimination. Therefore, it was concluded that Fe3O4-ACLM was a suitable substitute for other adsorbents in the removal of CIP from aqueous solutions.