1 Introduction

Dye contaminants are of the main pollution sources of water and their presence can result in serious problems for many forms of life. These toxic substances have been utilized in different industries such as printing, paper, textile, food, cosmetics and electroplating [1]. Considerable amounts of wastewater are generated during the utilization of dyes in the industrial processes [2]. Therefore, improper disposal of these contaminants can bring about serious environmental problems [3].

Methylene blue (MB), one of these toxic dyes, is widely used in the large number of industries as a colorant agent. Although this compound is not classified as so hazardous material, it can cause permanent eye damages in human and animals. On inhalation, it can lead to short period of rapid or difficult breathing. Ingestion of methylene blue results in the burning sensation and may give rise to several problems such as nausea, vomiting, profuse sweating, mental confusion, and methemoglobinemia [1,2,3]. Therefore, removal of this harmful organic component from the wastewaters is of great necessity [4].

Several processes such as trickling filter [5], membrane filtration [6], activated sludge [7], chemical coagulation, and adsorption [8, 9] have been developed for removing of dye contaminants from wastewaters. Among them, adsorption has been paid more attentions due to the effectiveness, low cost, facile process, relatively low cost operating factors, and also the possibility of the recycling process [9]. Different adsorbents such as natural zeolites [10], mesoporous SiO2 [11, 12], metal organic frameworks (MOFs) [13], graphene-based materials [14], and artificially [15, 16] and naturally (biomass and waste material) derived porous carbons [17, 18] have been developed for this purpose. Natural precursor–derived porous carbons which are cost-effective can be obtained by physical (with CO2 or steam) or chemical activation of the carbon precursors. Chemical approaches were performed with activating agents such as NaOH, KOH, ZnCl2, and H3PO4 in order to induce micro- and mesopores in the carbon containing materials and remove some of the unfavorable contents of the char [19,20,21]. Porosities, especially the micro-porosities, increase the surface area of the prepared activated carbons. Carbon source, activating agent, activation temperature, and the ratio of activating agent to carbon source are the main parameters which affect the characteristics of the porosities, i.e., mean pore size, distribution, and total volume of the porosities. These, in turn, determine the surface area of the prepared activated carbons. The surface area of various activated carbons differs from several hundred to about 3000 m2 g−1 [22,23,24].

Applying activated carbons, the adsorption behavior of methylene blue has extensively been investigated both thermodynamically and kinetically [25, 26]. These studies revealed that the adsorption capacities of the activated carbons depend on several parameters such as porosity size (meso- and especially micropores), porosity distribution, total pore volume, and also on the heteroatoms (boron, nitrogen, oxygen, sulfur, and phosphor) contents of the adsorbents [27,28,29]. To the best of the authors’ knowledge, there is not any straight dependency between the specific surface area of the adsorbent and the adsorption capacity. Boron-doped biomass-derived porous carbon with moderate surface area amount of about 650 m2 g−1 shows high adsorption capacity of about 2440 mg g−1 [27] and on the other hand, activated carbons with very high surface area amounts of 1534, 1937, 2086 and 2573 m2 g−1 reveal moderate amount of adsorption capacity (704, 916, 385 and 570 mg g−1, respectively) [30,31,32,33]. The applicability of several adsorption theories has also been studied for the adsorption process and it has been found that Langmuir adsorption theory (monolayer chemical adsorption of the adsorbate on the activated carbon) is the most acceptable theory [34,35,36]. This finding has also been approved by the pseudo-second-order kinetic theory [37, 38]. Thermodynamic studies revealed that the adsorption process of methylene blue on porous carbons, like namely all other dyes, is endothermic. Also, the entropy of the adsorption reaction will be increased during the dye removal process [16, 35, 39].

Although there are numerous studies about the adsorption behavior of porous carbons, to the best of the author’s knowledge, there is no any characteristic view about the adsorption process regarding pre- and post-adsorption process. In other words, there is no any deep structural insight on the adsorption process. The present study focuses on the relation between the adsorption behaviors of Azolla-derived porous carbons and their structural characteristics before and after the absorption process. The experimental data have been collected from the adsorption of methylene blue on the relatively high surface area Azolla-derived porous carbons. Based on the fitness results, the best isotherm and also the adsorption energy of methylene blue on Azolla-derived activated carbon were determined. Following this, pseudo-first- and pseudo-second-order kinetics model were applied to better understand the adsorption process. For better insight about the structural effects of the adsorption process, XRD, FTIR, Raman, and XPS analyses have been conducted on the prepared activated carbons before and after the adsorption process.

2 Experimental

2.1 Materials

All chemicals and reagents were of analytical grade and used as received without any further purification. Azolla biomass was collected from Anzali Lagoon. In order to remove any contaminants from this biomass, it was washed several times before treatment. Sodium hydroxide was purchased from Merck and nitric acid (37%) (HNO3) and also methylene blue was purchased from Dr. Mojallali™ Company.

2.2 Synthesis of activated carbon

Porous activated carbons have been prepared using an alkaline approach which is categorized as a chemical activation process. At first, Azolla biomass was pre-carbonized for 3 h at 300 °C under argon atmosphere. Obtained biochar was mixed with sodium hydroxide (NaOH) with the weight ratio of 1 to 3 (biochar to NaOH). Mixing treatment was continued until homogenization. The obtained mixture was pyrolyzed for 3 h at 600 °C in a tubular electric furnace, under argon atmosphere. The black product was washed with nitric acid solution (1 M) to remove the activation agents and other impurities. This treatment was followed by washing using deionized water until neutral pH, (Az-C).

3 Characterization

PXRD patterns were collected on a Philips (Holland) PW1730 diffractometer with Cu Kα (λ = 1.54 Å). N2 adsorption–desorption isotherm was conducted on a BELSORP mini II equipment at 77 K, degassing time and temperature of which are 6 h and 150 °C, respectively. Raman spectroscopy was conducted on a Takram micro-Raman spectrometer (Teksan™, Iran). FTIR spectroscopy was accomplished on a Shimadzo 8400s Spectrometer. X-ray photoelectron spectroscopy (XPS) measurements were carried out on an Al Ka source (XPS Spectrometer Kratos AXIS Supra). High-resolution transmission electron micrographs were obtained on a 300-kV FEI (US) TITAN microscope (HR-TEM). The absorbance of MB solution was measured by an UV-1700 (Shimadzu, Japan) UV–visible spectrophotometer.

3.1 Adsorption of methylene blue

Kinetics and thermodynamics of MB adsorption on the Azolla-derived porous carbon have been investigated during the adsorption process. In a typical process, 5 mg of porous carbon powders was added to different volumes (from 10 to 40 ml) of MB aqueous solutions. The concentration of all the solutions was fixed at 400 mg l−1. The adsorption processes were continued until equilibrium establishment (up to 45 min). All the experiments were conducted at ambient temperature.

Using the following equation, the adsorption capacity of each experiment at every adsorption time was determined [40]:

$$ {Q}_{\mathrm{t}}=\left({C}_0-{C}_{\mathrm{t}}\right)\ \frac{V}{m} $$
(1)

In this expression, Qt (mg g−1) is the time-dependent adsorbed amount of dyes, C0 and Ct (mg l−1) are the solute concentrations at the beginning and after the selected time, respectively, V (ml) is the volume of dye solution which was fixed at 20 ml, and m (mg) is the mass of the porous activated carbon.

The fitness of four well-known adsorption isotherms models, i.e., Langmuir, Freundlich, Temkin and Dubinin–Radushkevich, has been investigated for the obtained equilibrium solution concentrations. Parameters including in these isotherms have been introduced as follows:

  1. a)

    Langmuir’s isotherm:

Langmuir isotherm which is obeyed in systems with homogeneous interaction between the adsorbent and the adsorbate can be introduced as:

$$ {Q}_{\mathrm{e}}=\frac{k_{\mathrm{l}}{Q}_{\mathrm{m}}{C}_{\mathrm{e}}}{1+{k}_{\mathrm{l}}{C}_{\mathrm{e}}} $$
(2)

In the above equation, Ce (mg l−1) is the equilibrium concentration of the solution, Qe (mg g−1) is the amount of adsorbed dye when equilibrium condition was established, Qm (mg g−1) is the maximum adsorption capacity, and KL (l mg−1) is the Langmuir constant [16].

  1. b)

    Freundlich isotherm:

Freundlich isotherm is followed in heterogeneous adsorption systems. This isotherm can be written as:

$$ {Q}_{\mathrm{e}}={K}_{\mathrm{F}}{C}_{\mathrm{e}}^{\frac{1}{n}} $$
(3)

KF (mg g−1) and n are constant coefficients and relate to the absorption capacity and absorption intensity, respectively [16].

  1. c)

    Temkin isotherm:

Temkin isotherm equation which based on heterogeneous adsorption behavior is as follows:

$$ {Q}_e=\frac{\mathrm{RT}}{b_{\mathrm{T}}}\ln \left({K}_{\mathrm{T}}\right)+\frac{\mathrm{RT}}{b_{\mathrm{T}}}\ln \left({C}_e\right) $$
(4)

where KT and bT are the equilibrium binding (l mg−1) and the Temkin isotherm constants, respectively.

Dubinin–Radushkevich (D–R) isotherm is as follows.

This isotherm, which is applied for heterogeneous adsorption systems, can be presented as:

$$ \ln {Q}_{\mathrm{e}}=\ln {Q}_{\mathrm{m}}-\beta {\epsilon}^2 $$
(5)
$$ \epsilon =\mathrm{RTln}\left(1+\frac{1}{C_{\mathrm{e}}}\right) $$
(6)
$$ {E}_{\mathrm{a}}=\frac{1}{\sqrt{2\beta }} $$
(7)

In these expressions, β (mol2 kJ−2) is D–R isotherm constant, ε is the Polanyi potential, and Ea (kJ mol−1) is the adsorption energy. T (k) is the adsorption temperature and also R with the value of 8.314 is the gas universal constant (J mol−1 k−1) [26].

Adsorption kinetics of the prepared activated carbon for methylene blue has also been investigated. Two adsorption kinetics models, namely pseudo-first- and pseudo-second-order, were applied to investigate their applicability on the experimental adsorption data.

In order to investigate the adsorption mechanism of methylene blue on the Azolla-derived activated carbon, XRD, FTIR, Raman, and XPS analyses have been conducted before and after the methylene blue adsorption on the prepared carbons. For this purpose, 30 mg Azolla-derived porous carbons was added to 20 and 40 ml of 400 mg l−1 of methylene blue solutions, (Az-C-20 and Az-C-40, respectively). After completion of the adsorption process and establishing of equilibrium, the carbon and adsorbed methylene blue mixtures were filtered and dried for further characterization. This adsorption process results in the addition of 8 and 16 mg of methylene blue into the 30 mg activated carbon.

4 Results and discussion

4.1 Equilibrium and kinetics of adsorption behavior

Equilibrium adsorption capacity results of MB on Az-C have been demonstrated in Fig. 1a. Linear forms of Langmuir, Freundlich, Temkin and Dubinin–Radushkevich [37, 41] isotherms were also included in Fig. 1b–e. The obtained data from these plots have been shown in Table 1. Comparison between these results reveals that there is a better consistency between the methylene blue adsorption behavior of Az-C with the Langmuir (homogeneous monolayer adsorption) model. Therefore, it can be concluded that all the adsorption sites of the prepared activated carbon have similar adsorption conditions (both energetically and availability). This consistency also implies that the functional groups were homogeneously distributed on the surface of Az-C. Considering the adsorption process, both in aqueous and gaseous atmospheres, Langmuir is more common model which is followed with the adsorbent–adsorbate systems [15, 16]. Based on the Langmuir isotherm, the adsorption capacity of Azolla-derived porous carbon is about 1930 mg g−1 which is one of the highest adsorption capacity of porous activated carbon for methylene blue [15, 42]. This high value is attributed to both surface physical (high specific surface area, meso- and micro-porosities) and chemical (nitrogen-containing functional groups) characteristics of the activated carbon [42]. The equilibrium parameter (RL) can be calculated based on Eq. 8:

$$ {R}_{\mathrm{L}}=\frac{1}{1+{K}_{\mathrm{L}}{C}_{\mathrm{e}}} $$
(8)
Fig. 1
figure 1

a Adsorption capacity of methylene blue on Azolla-derived porous carbon (Az-C) in concomitant with the Langmuir and Freundlich plots of the experimental results, b linearized forms of Langmuir, c Freundlich, d Temkin, and e Dubinin–Radushkevich (D–R) isotherms for the adsorption of methylene blue on Az-C

Full size image
Table 1 Langmuir, Freundlich, Temkin, and D–R isotherms data for the adsorption of methylene blue on the Azolla-derived porous carbon
Full size table

RL is about 0.00003 which is approximately equal to zero and therefore implies irreversible adsorption of methylene blue on Az-C [43]. Monolayer and limiting value of methylene blue adsorption and also very low equilibrium parameter predict the chemical adsorption mechanism of this adsorbate on Az-C. This prediction can further be approved by the mean adsorption energy. From expression 7, the mean adsorption energy of methylene blue on Az-C is 21.76 kJ mol−1 which is so greater than 8 kJ mol−1 which is the top limit for physical adsorption. This finding emphasizes on the strong chemical interaction between the adsorbent and the adsorbate [38, 44].

The kinetic behavior of Az-C aqueous solution of methylene blue system was also investigated for initial methylene blue concentration of 400 mg l−1. The kinetic plots of this system are presented in Fig. 2a–f. From Fig. 2a, b, it can be observed that after 15 min about 95% of methylene blue was adsorbed on Az-C and the equilibrium condition is established after about 45 min. The high adsorption rate of methylene blue on Az-C implies high driving force for interaction between dye molecules and the adsorbent. For better insight into the kinetics of the adsorption process, pseudo-first- and pseudo-second-order kinetic models were applied to the experimental results. These two models can be expressed as follows, respectively [9, 38, 44]:

$$ \ln \left({q}_e-{q}_t\right)=\ln \left({q}_e\right)-{k}_1t $$
(9)
$$ \frac{t}{q_t}=\frac{1}{k_2{q}_e^2}+\frac{t}{q_e} $$
(10)
Fig. 2
figure 2

Kinetic plots of methylene blue adsorption on Az-C. a Methylene blue concentration versus the time of adsorption. b Adsorption capacity versus the time of adsorption. c Pseudo-first- and d pseudo-second-order model of adsorption. e, f Intra-particle and Boyd models of transport

Full size image

The results of these analyses can be followed in Fig. 2c, d. Better agreement between the experimental results and the pseudo-second-order kinetic model can be observed. This finding emphasizes that the adsorption process is chemical. The same behavior has been proposed for the adsorption of methylene blue on the several other adsorbents [37, 38, 44, 45]. For better insight on the rate-limiting step [external transport (film diffusion) or intra-particle diffusion], two other models (intra-particle diffusion and Boyd diffusion) were fitted to the experimental results. Intra-particle diffusion model is described as follows:

$$ {q}_{\mathrm{t}}={k}_{\mathrm{i}}{t}^{0.5}+C $$
(11)

where qt (mg g−1) is the adsorption capacity at time t, ki (mg g−1 min−0.5) is the rate constant of the intra-particle diffusion model and C is a constant [44]. Two parameters of this model, i.e., ki and C, can easily be determined by the slope and intercept of qt versus t0.5 [9, 45]. If the graph pass through the origin and also has a good fitting with the experimental data points (R2~1), the rate-limiting step in the adsorption process is intra-particle diffusion of the dye ions. From Fig. 2e, it is clear that despite fair fitting of the experimental data points with the model, the results did not pass through the origin; therefore, at this point, one cannot decide for the intra-particle diffusion as the rate controlling step and Boyd model should be examined. This model is described as:

$$ Bt=-0.4977-\ln \left(1-F\right) $$
(12)
$$ F=\frac{q_{\mathrm{t}}}{q_{\mathrm{e}}} $$
(13)

The plot of Bt versus t is called the Boyd plot. If a straight line passing the origin with high R2 value is achieved during this data fitting, intra-particle diffusion is the rate-limiting step [9]. From the Boyd plot of methylene blue adsorption on Az-C, Fig. 2f, this is the case for this adsorption process. Little deviation of the plot from the origin (0.07) seems negligible. Considering the high adsorption capacity, chemical nature of the adsorption process, and intra-particle diffusion of the adsorbate as the rate-limiting step, it can be concluded that physical and chemical characteristics of the adsorbent and their variations during the adsorption process are so determinative.

4.2 Structural interpretation of the adsorption mechanism

The HR-TEM micrograph of Az-C in concomitant with N2 adsorption–desorption isotherm, BJH, and micropore (MP) plots is presented in Fig. 3a–d. Worm-like porous structure which comprises of large amounts of micropores can be observed in the prepared Azolla-derived porous carbons (Fig. 3a). Total surface area of this porous structure is about 1437.6 m2 g−1 which was obtained from BET analysis of the adsorption–desorption data. Also, the total pore volume of this structure is 0.74 cm3 g−1 (Fig. 3b). Comparing Fig. 3c and d, it will be determined that, in consistence with Fig. 3a, micropores have greater amount than meso- and macropores. Micropore surface area and pore volume are 1352.2 m2 g−1 and 0.599 cm3 g−1, respectively. In fact, approximately 94% of the surface area and 80% of the pore volume of this biomass-derived carbon come from micropores. High surface area of this porous carbon is a prerequisite for high adsorption capacity, especially for Langmuir model of adsorption. In addition, intra-particle diffusion as the limiting rate step for MB adsorption (Fig. 2f) [42] can be attributed to the meso-sized porosities which facilities the diffusion of cationic dye ions and to the high density of micropores with diameters less than 1 nm (Fig. 3c).

Fig. 3
figure 3

a The HR-TEM micrograph of the Azolla-derived porous carbon (Az-C). b N2 sorption isotherm. c, d BJH and MP plots of the Azolla-derived porous carbon, respectively

Full size image

XRD, Raman, and FTIR spectra of Az-C before and after adsorption of MB has been demonstrated in Fig. 4a–c. Methylene blue adsorption on the prepared activated carbon resulted in a shift of (002) crystallographic plane of the turbostratic carbon from about 22 to about 25° (Fig. 4a) [46]. This shift to higher 2θ can be attributed to the induction of compressive stress on the graphite crystallites. Chemical interaction of methylene blue molecules with the surface functional groups and atoms of Az-C and also their diffusion into the micro-porosities of Az-C can be considered as the reason for this compressive stress. A blue shift of X-ray diffraction patterns of the adsorbed porous carbons to higher 2θ also implies that no intercalation of methylene blue molecules occurred into the (002) planes. This may be attributed to the low value of the adsorbed methylene blue ions on the porous carbon [47]. In fact, several mechanisms may be active during the adsorption process of MB from the contaminated water. These are adsorption of cationic ions on the surface functional groups by physical or chemical interactions, entrapping of them in to the micro-porosities and intercalation of MB ions between the crystal planes (absorption). If intercalation was the active mechanism, it would induce tensile stress on the crystallographic planes of the porous carbon. This tensile stress, in turn, would lead to red shift in the characteristic peaks of porous carbon [48]. Considering the blue shift of the characteristic peaks and also the same trend observed during water decolorization (Fig. 2), it can be concluded that intercalation (and absorption) is not the prevailing mechanism of water decolorization by Azolla-derived porous carbon. From Raman spectra of Az-C before and after MB adsorption (Fig. 4b and inset), it is observed that graphite band (G-band) is blue shifted (from 1593 to 1597.75 cm−1) which means a decrease in C = C graphite band length (compressive stress on graphite structure) which is in agreement with the XRD results. At the same time, the relative graphitization value of the carbon structure (\( \frac{I_D}{I_G}\Big) \) increases from 0.88 for Az-C to 0.93 for Az-C-20 and 0.97 for Az-C-40. This increase in relative graphitization value means that adsorption of methylene blue will result in the increase in the defects of the carbon structure. Comparing the FTIR spectra of Az-C and methylene-infiltrated Az-C (Az-C-20 and Az-C-40), it can be deduced that there is two distinct shift in the location of characteristic peaks of Az-C, which are shown with arrows in Fig. 4c. The first shift is occurred for the peaks attributed to C-O bands (1050 cm−1) [30]. Increasing the amount of the adsorbed methylene blue increases the wavenumber of the characteristic peak of C-O band to about 1160 cm−1. The intensity of this peak decreases during the adsorption process. Also, the characteristic peaks of C=C and N-H bands at around 1570–1650 cm−1 [49] shifts to higher wavenumbers and also, the intensity of this peak increases. Other characteristic peaks of Az-C seem without any notable changes during adsorption process. The change in the intensity and the wavenumber of these functional groups emphasizes the involvement of C-O, N-H, and C=C groups on adsorption process [9, 38]. Considering these observed shifts in the characteristic peaks of some functional groups (in FTIR analysis) and crystal plains (in XRD and Raman analyses) emphasizes on the strong electrostatic attraction or chemical bonding between a (some) group(s) of the adsorbed dye and a (some) functional group(s) of the adsorbent.

Fig. 4
figure 4

a XRD, b Raman, and c FTIR spectra of Az-C before and after adsorption of methylene blue

Full size image

XPS results can give better insight about the surface chemistry of the prepared porous carbon and also chemical interaction between the methylene blue and the adsorbent. Deconvoluted quantitative XPS results of C1s, N1s, O1s, and S2p spectra of Azolla-derived porous carbon before and after the adsorption process have been shown in Table 2. Except with some difference between C, O, and S surface functionalities of Az-C before and after the adsorption treatment, it seems that nitrogen surface chemistry is substantially different in these three specimens. N1s spectra of Azolla-derived porous carbon before and after adsorption of methylene blue have been demonstrated in Fig. 5. It is clear that, as a result of the adsorption process, pyrrolic nitrogen group diminished and N-(C)3 and H-N-(C)2 nitrogen groups peak appeared. During the adsorption process, the relative percent of this new peak increased from 39.2 to 79.7 which is in agreement with the amounts of the adsorbed methylene blue. Therefore, this peak can be attributed to the nitrogen groups in the adsorbed dye. In addition and by continuation of the adsorption process, graphitic nitrogen also diminished. The relative percent of pyridinic nitrogen group also decreased. But this decrease is not as high as that of pyrrolic and graphitic nitrogen groups. Therefore, it can be inferred that ionized nitrogen atoms of methylene blue which have positive electrical charge and a vacant 2p orbital, chemically, interact with pyrrolic, graphitic, and to some extent with pyridinic nitrogen surface sites of Az-C. It seems that the adsorbed dye works like a coating and prevents the detection of these functional groups. Elaborating the detailed mechanism of this interaction is out of the scope of the present research.

Table 2 Physicochemical changes of surface bonds in Az-C before and after adsorption of methylene blue, extracted from XPS results
Full size table
Fig. 5
figure 5

Deconvoluted XPS N1s spectra of Azolla-derived porous carbon before (Az-C) and after (Az-C-20, Az-C-40) adsorption process

Full size image

5 Conclusion

In the present research, equilibrium, kinetics, and structural changes during the adsorption of methylene blue on a relatively large surface area (1437 m2 g−1) Azolla-derived porous carbon was investigated. Several adsorption theories, Langmuir, Freundlich, Temkin, and Dubinin–Radushkevich were investigated for the adsorption process of methylene blue on the prepared porous carbon. These considerations were declared that Langmuir (monolayer, chemical, and homogeneous) adsorption theory is consistence with the experimental results with high adsorption capacity of about 1930 mg g−1. Adsorption energy of methylene on the prepared porous carbon was determined to be about 21.76 kJ mol−1 which further confirm the chemical adsorption. From kinetics point of view, pseudo-second-order kinetic model is dominant in this adsorption process. This finding further emphasizes on chemical nature of the adsorption process. Applying intra-particle and Boyd diffusion models, it was determined that the rate-limiting step in this process is the diffusion of methylene blue in the meso- and micro-porosities of the prepared carbons. Based on X-ray diffraction results before and after the adsorption process, there is no any sign (induction of tensile stress in the graphitic planes) for intercalation of methylene blue in the inter-crystalline plane spaces of the porous carbon. Adsorption process results in the blue shift of X-ray diffraction peak of (002) graphitic planes from about 22 to about 25° which means the induction of compressive stresses on the graphitic planes. This compressive stresses come from the diffusion of methylene blue in to the meso- and micro-porosities of the porous carbon structure. Adsorption of methylene blue leads to the reduction of relative graphitization of the prepared carbon. From FTIR and XPS results, it was declared that nitrogen functional groups are the main functional groups which involve in the adsorption process of methylene blue. During the adsorption process, the ratio of pyridinic nitrogen groups decreased and pyrrolic and graphitic nitrogen groups eliminated from the surface of Azolla-derived porous carbon. At the same time, two new functional nitrogen groups, namely and N-(C)3 and H-N-(C)2, appeared in the surface of the prepared adsorbent.