1 Introduction

Levulinic acid contains a ketone carbonyl group and an acidic carboxyl group and it is called as keto acid. Levulinic acid is a remarkable versatile acid with a wide range of uses in various areas. It is utilized as a coating material, antifreeze, plasticizer, solvent, stabilizer, food flavoring, and animal feed. Furthermore, it has many uses in the polymer, coating, printing, textile, and pharmaceutical industries as well as in the manufacturing of nylon and rubber, glass-like synthetic fibers and resins, and hydraulic brake fluids [1].

Levulinic acid is mostly produced by the hydrothermal degradation of carbohydrate-containing materials. Firstly, the carbohydrate source is hydrolyzed to form hydroxymethyl furfural in the presence of diluted mineral acid. In the next stage, hydroxymethyl furfural is hydrolyzed and levulinic acid occurs. Furthermore, by-products can form during this process and acid production efficiency decreases with by-product formation. So, the cost and complexity of the acid purification and recovery process increase [1]. Thereby, the by-products require to be removed from the final product. However, levulinic acid can be also produced by fermentation of sucrose that is a biotechnological production method of great importance, and it is obtained in the form of aqueous solutions during the fermentation [2, 3]. The obtained acid should be removed from the water to carry out the fermentation process continuously [4]. The purification step of the manufacturing process is important to decrease the production costs and increase energy efficiency and improve economic performance and environmental impact [5, 6]. Hence, it is substantial to remove and purify levulinic acid during the manufacturing processes as well as from aqueous media or wastewater formed in the use of the chemical industries.

A lot of separation techniques, including extraction, adsorption, ion-exchange, precipitation, ultrafiltration, nanofiltration, electrodialysis, reverse osmosis, membrane separation, and distillation have been extensively employed to remove carboxylic acids from water [7,8,9]. Among these techniques, extraction [10], nanofiltration and electrodialysis [11], and adsorption [12, 13] have been employed to remove levulinic acid from aqueous solutions. Adsorption is an effective process, because it has substantial features, like high separation efficiency, quick response, simplicity, absence of sludge, low instrumentation and operation costs, and applicability of diverse adsorbents [14, 15].

Various adsorbents such as activated carbon [16], Amberlite XAD-4 [12], Amberlite IRA-67 [17], basic polymeric adsorbents (335 resin, D301, D315) [13, 18, 19], and microporous hyper-cross-linked polymers [20, 21] have been investigated in the adsorption of levulinic acid. It is noted that the adsorption studies of levulinic acid were limited. Recently, the usage of nano-adsorbents in adsorption processes has gained importance since they exhibit better adsorption properties [22]. Especially, there is a growing demand for testing nano-adsorbents in the removal of organic compounds from aqueous solutions. Clays are natural, environmentally friendly substances with a high specific surface area and are extensively preferred for the adsorption and removal of various substances [23, 24]. Besides, they have a low-cost compared with conventional adsorbents. The most commonly used clay as nano-adsorbent is montmorillonite. Montmorillonite is a kind of clay material that is abundant in nature [25]. Nano-adsorbent Cloisite 20A clay is montmorillonite modified by adding quaternary ammonium salts. Based on these properties of clays, in this study, montmorillonite and Cloisite 20A clays are considered as novel adsorbents for the uptake of levulinic acid from aqueous solutions because montmorillonite and Cloisite 20A (modified montmorillonite) as clay nano-adsorbents have not been examined in levulinic acid adsorption so far. Hence, in this adsorption research, clay nano-adsorbents were tested and compared for the adsorption of levulinic acid. For this purpose, the effects of various adsorption factors like contact time, initial acid concentration, adsorbent amount, and temperature were investigated. So, the equilibrium, kinetic, and thermodynamic data were obtained to characterize the adsorption process.

2 Materials and method

Aqueous acid solutions with different concentrations were prepared by dissolution of levulinic acid (> 98%) in distilled water. Levulinic acid was supplied from ACROS Organics. Montmorillonite (98%) and Cloisite 20A (> 97%) were obtained from Southern Clay Products (Texas, U.S.A.). Montmorillonite is a hydrated aluminum silicate with sodium as the dominant cation exchanger. Cloisite 20A is montmorillonite modified by adding dimethyl dihydrogenated tallow quaternary ammonium chloride. Tallow is derived from beef and is mainly consists of 18 carbon chains (approximately 65% C18, 30% C16, and 5% C14). Hydrogenated tallow is obtained from tallow by the hydrogenation of double bonds. Cloisite 20A consists of 2 tallow groups; hence, it is more hydrophobic [26]. The clays were used without further treatment. The main characteristic properties of clays provided by the supplier were also summarized in Table 1. The operating ranges of adsorption parameters such as initial acid concentration, adsorbent amounts, and temperature were determined from the preliminary experiments. So, batch adsorption experiments were exerted by shaking glass Erlenmeyer flasks containing 5 mL levulinic acid solutions (concentrations of 20–100 g L−1) and adsorbents (amounts of 0.05–0.25 g) in a water bath for the determined equilibrium time (180 min) at different temperatures (25–45 °C). The rate of shaker (Nüve ST 30) was 150 rpm. After adsorption operation, the samples were filtered and the adsorbents were separated from the aqueous solutions by utilizing the filtration procedure. The levulinic acid concentrations before and after the adsorption operation were detected by using SI Analytics automatic titrator using 0.1 N NaOH.

Table 1 The characteristic properties of clays
Full size table

Adsorption capacity (qe, mg g−1) of adsorbents at equilibrium for levulinic acid adsorption were calculated from the following Eq. (1):

$$ {q}_e=\left[\frac{\left({C}_0-{C}_e\right)}{M}\right]\ast V $$
(1)

where C0 signifies the initial acid concentration (mg L−1); Ce denotes the equilibrium acid concentration (mg L−1); V represents the volume of solution (L); and M symbolizes the quantity of adsorbents (g).

3 Results and discussion

3.1 Determination of equilibrium adsorption time

To determine the equilibrium time, the batch experiments were exerted using an initial concentration of levulinic acid solution of 100 g L−1 and a clay amount of 0.05 g. Figure 1 shows the variation of levulinic acid adsorption over time for both adsorbents. As can be observed from Fig. 1, the levulinic acid adsorption increased as time increased in the initial stages; then, the adsorption slowed until equilibrium was attained because, initially, adsorption is quick owing to the free adsorption zones on the adsorbent surface. The adsorption then slows down because of the reduction in the number of adsorption sites present. Considering the values of the adsorption capacity of both adsorbents depicted in Fig. 1, it was seen that the adsorption had reached the equilibrium between 150 and 180 min. So, the equilibrium adsorption time was determined as 180 min.

Fig. 1
figure 1

The variation of levulinic acid adsorption over time for montmorillonite and Cloisite 20A. Operating conditions: C0 = 100 g L−1, M = 0.05 g, and T = 25 °C

Full size image

3.2 The effect of initial acid concentration

In order to examine the effect of initial acid concentration on adsorption, the experiments were exerted using five concentrations of levulinic acid of about 20, 40, 60, 80, and 100 g L−1 with the period of equilibrium adsorption time at 25 °C. The adsorbent amount utilized was 0.05 g. The results of these experiments were plotted in Fig. 2. When Fig. 2 was examined, it can be seen that the values of adsorption capacity increased from 152.70 to 355.69 mg g−1 for montmorillonite and from 185.99 to 428.63 mg g−1 for Cloisite 20A with the increasing the levulinic acid concentration. Uslu et al. [17] separated levulinic acid from its aqueous solutions by Amberlite IRA-67 and they also found that the adsorption capacity increased with the increased concentration of acid. Similarly, Datta and Uslu [12] investigated the levulinic acid adsorption with Amberlite XAD-4. They reported that the values of adsorption capacity increased with an increase in the initial concentration of levulinic acid in the solution.

Fig. 2
figure 2

The effect of initial acid concentration on levulinic acid adsorption by montmorillonite and Cloisite 20A. Operating conditions: M = 0.05 g, T = 25 °C, and t = 180 min

Full size image

3.3 Effect of adsorbent amount

The effect of the adsorbent amount was examined by using various amounts varied between 0.05 and 0.25 g at the initial acid concentration of 100 g L−1 at 25 °C. The effect of adsorbent amount on the adsorption capacity for levulinic acid was depicted in Fig. 3. It was seen from Fig. 3 that the values of adsorption capacity decreased with the increase in the amount of nano-adsorbents. The maximum adsorption capacity values were obtained as 428.63 mg g−1 for Cloisite 20A and 355.69 mg g−1 for montmorillonite in case of using an amount of 0.05 g.

Fig. 3
figure 3

The effect of amounts of montmorillonite and Cloisite 20A on levulinic acid adsorption. Operating conditions: C0 = 100 g L−1, T = 25 °C, and t = 180 min

Full size image

3.4 Effect of temperature

To observe the effect of temperature on levulinic acid adsorption by montmorillonite and Cloisite 20A, the experiments were exerted at three dissimilar temperatures (25, 35, and 45 °C). The initial levulinic acid concentration of about 100 g L−1 and the adsorbent amount of 0.05 g were used. The acquired results were demonstrated in Fig. 4. As can be seen from Fig. 4, the values of adsorption capacity were decreased with an increase in temperature, and the trends were obtained as same for both adsorbents. It was clear from these results that the adsorption of levulinic acid by montmorillonite and Cloisite 20A was exothermic. This result was confirmed by observations from similar studies of levulinic acid [12, 17].

Fig. 4
figure 4

The effect of temperature on levulinic acid adsorption by montmorillonite and Cloisite 20A. Operating conditions: C0 = 100 g L−1, M = 0.05 g, and t = 180 min

Full size image

3.5 Equilibrium adsorption models

The equilibrium data were appraised by utilizing the adsorption isotherm models including Langmuir, Freundlich, and Temkin isotherm models. The isotherm models and equations were given in Table 2 [27, 28].

Table 2 Adsorption isotherm models
Full size table

Figures 5 and 6 show the linear plots of Langmuir, Freundlich, and Temkin isotherm models for levulinic acid adsorption by montmorillonite and Cloisite 20A, respectively. Table 3 indicates Langmuir, Freundlich, and Temkin isotherm model parameters and regression coefficients calculated from these illustrated linear plots. When the values of regression coefficients given in Table 3 were examined, the Freundlich isotherm model exhibited the best fit to the equilibrium adsorption data of both adsorbents. The Freundlich isotherm model approach assumes that the adsorption occurs multilayer on the adsorbent surface and implies that the adsorbent surface is heterogeneous [22].

Fig. 5
figure 5

Linear fits of a Langmuir, b Freundlich, and c Temkin isotherm models for the levulinic acid adsorption by montmorillonite

Full size image
Fig. 6
figure 6

Linear fits of a Langmuir, b Freundlich, and c Temkin isotherm models for the levulinic acid adsorption by Cloisite 20A

Full size image
Table 3 Isotherm parameters and regression coefficients for levulinic acid adsorption by montmorillonite and Cloisite 20A
Full size table

3.6 Adsorption kinetic models

The adsorption kinetic of levulinic acid by montmorillonite and Cloisite 20A were examined by utilizing three kinetic models including pseudo-first-order, pseudo-second-order, and Elovich kinetic models. The kinetic models and equations were given in Table 4 [29, 30]. The kinetic experiments were actualized at different contact time intervals (30–180 min) with initial levulinic acid concentration of 100 g L−1, nano-clay amount of 0.05 g, and temperature of 25 °C.

Table 4 Adsorption kinetic models and equations
Full size table

Figures 7 and 8 demonstrate the linear plots of pseudo-first-order, pseudo-second-order, and Elovich kinetic models for levulinic acid adsorption by montmorillonite and Cloisite 20A, respectively. Table 5 shows the kinetic model parameters and regression coefficients calculated from these depicted linear plots. When the values of the regression coefficient shown in Table 5 were analyzed, the pseudo-second-order kinetic model exhibited the best fit to the kinetic adsorption data for both adsorbents. Pseudo-second-order kinetic model implies that the adsorption mechanism depends on the adsorbent and adsorbate [31, 32].

Fig. 7
figure 7

Linear fits of a pseudo-first-order, b pseudo-second-order, and c Elovich kinetic models for levulinic acid adsorption by montmorillonite

Full size image
Fig. 8
figure 8

Linear fits of a pseudo-first-order, b pseudo-second-order, and c Elovich kinetic models for levulinic acid adsorption by Cloisite 20A

Full size image
Table 5 Kinetic model parameters and regression coefficients for levulinic acid adsorption by montmorillonite and Cloisite 20A
Full size table

3.7 Adsorption thermodynamic

To investigate levulinic acid adsorption thermodynamic, three different temperatures (25, 35, and 45 °C) were studied at the initial levulinic acid concentration of 100 g L−1 and adsorbent amount of 0.05 g. The thermodynamic parameters including Gibb’s free energy (ΔG0), enthalpy (ΔH0), and entropy (ΔS0) were determined. ΔG0, ΔH0, and ΔS0 were calculated by means of equations given in Table 6 [33, 34].

Table 6 Adsorption thermodynamic equations
Full size table

Figure 9 shows the linear plots of \( \ln \left({K}_D\right)\ \mathrm{vs}.\kern0.5em \frac{1\ }{T} \) for levulinic acid adsorption by montmorillonite and Cloisite 20A. Table 7 shows thermodynamic parameters (ΔS0 and ΔH0) calculated from the intercept and slope of linear plots. As can be seen from Table 7, the adsorption of levulinic acid was exothermic (ΔH0 < 0) for both nano-adsorbents. It was also found ΔS0 < 0 indicating that there was a decreased randomness at the adsorbent-adsorbate interface during the adsorption process [34]. In addition, the values of ΔG0 based on different temperatures were obtained as shown in Table 8. ΔG0 < 0 implies that adsorption is spontaneous and favorable [34]. Thus, the adsorption of levulinic acid by montmorillonite and Cloisite 20A occurs spontaneously and it is favorable (Cloisite 20A > montmorillonite).

Fig. 9
figure 9

Linear plots of \( \ln \left({K}_D\right)\ \mathrm{vs}.\kern0.5em \frac{1\ }{T}\kern0.75em \) for levulinic acid adsorption by a montmorillonite and b Cloisite 20A

Full size image
Table 7 Thermodynamic parameters and regression coefficients for levulinic acid adsorption by montmorillonite and Cloisite 20A
Full size table
Table 8 Gibb’s free energy (ΔG0) for levulinic acid adsorption by montmorillonite and Cloisite 20A
Full size table

3.8 Comparison of adsorption capacities

The abovementioned studies indicated that Cloisite 20A (modified montmorillonite) exhibited a higher selectivity and adsorption capacity than montmorillonite. The difference in the adsorption performance of absorbents can be attributed to the differences in the surface features of the adsorbents. This conclusion has proved that a surface modifier has a significant impact on the adsorption capacity. The modified montmorillonite surface is preferred due to the presence of quaternary ammonium chloride (cationic compound containing alkyl groups in long carbon chains) compared with the unmodified montmorillonite surface and also modified montmorillonite is more hydrophobic [33]. A clay mineral predominantly performs the adsorption with a cation exchange mechanism. The cations can be bonded with electrostatic forces into the surface of adsorbents [15]. In this context, the adsorption phenomenon of levulinic acid onto clays may be explained by the cation exchange mechanism. Table 9 demonstrates a comparison of the adsorption capacity of nano-adsorbents in the present study with the values of the adsorption capacity of various adsorbents reported in the literature for levulinic acid. As shown in Table 9, the adsorption capacity value obtained in this study was quite high compared with other reported capacity values. These results confirmed that montmorillonite and Cloisite 20A are efficient nano-adsorbents for the removal of levulinic acid from aqueous solutions.

Table 9 Adsorption capacity values of various adsorbents in the literature for levulinic acid
Full size table

4 Conclusion

The levulinic acid adsorption was tested by utilizing montmorillonite and Cloisite 20A (modified montmorillonite). It was found that these clay nano-adsorbents were effective for the removal of levulinic acid from aqueous solutions. It was also obtained that nano-clay Cloisite 20A exhibited a higher selectivity and adsorption capacity than montmorillonite. The equilibrium, kinetic, and thermodynamic studies were carried out for both adsorbents by investigating the effect of time, acid concentration, and temperature on the adsorption process. The adsorption capacities of adsorbents were increased as the initial levulinic acid concentration increased. The maximum adsorption capacities obtained 428.63 mg g−1 for Cloisite 20A and 355.69 mg g−1 for montmorillonite. The optimal adsorption operating conditions were determined as the contact adsorption time of 180 min, initial levulinic acid concentration of 100 g L−1, the amount of adsorbent of 0.05 g, and the temperature of 25 °C. The equilibrium data for two adsorbents were best described by the Freundlich isotherm model, demonstrating the multilayer adsorption process. The kinetic data showed a good fit by the pseudo-second-order kinetic model for two adsorbents. From the thermodynamic analysis, the adsorption of levulinic acid on the clay nano-adsorbents was found to be exothermic process. This study indicated that these clay nano-adsorbents are effective and can be used in the removal of levulinic acid from aqueous solutions. Besides, this study has suggested that these adsorbents can be used for the removal of other organic acids from the fermentation broths or mixture or waste solutions.