Introduction

Organic dyes are widely used in various industries such as textile dyeing (Patil et al. 2010), printing (Seema et al. 2018), cosmetics (Guerra et al. 2018), paper making (Blus et al. 2014), etc. Dyes in wastewater are harmful to environment and cause health problems because of the toxicity and difficulty of degradation. So, the removal of dyes from wastewater is important. Many methods have been developed to remove dyes such as coagulation-flocculation (Panswed and Wongchaisuwan 1986), oxidation (Malik and Saha 2003), membrane separation (Ciardelli et al. 2001), electrochemical processes (de Paiva et al. 2018), and adsorption (Mall et al. 2005). Among them, adsorption is an effective process for the removal of dyes because of its low cost, easy technical access and simplicity of design (Wang and Peng 2010; Cheng et al. 2016; Zheng and Wang 2009; Zhou et al. 2011). Different types of adsorbents are commonly used to remove dyes such as activated carbon (Luo and Zhang 2009), clays (Almeida et al. 2009), natural zeolites (Wang et al. 2010), industrial byproducts and wastes (Ahmad et al. 2009), natural polymers (Annadurai et al. 2002), and hydrogels (Soleimani et al. 2018).

It was well-known that cross-linked polymeric materials especially hydrogels with functional groups such as carboxylic acid, hydroxyl, amine and sulfonic acid groups can be used as complexing agents for removal of dyes from aqueous solutions (Bekiari et al. 2008; Yetimoğlu et al. 2007; Rahchamani et al. 2011; Atta et al. 2012). When hydrogels encounter aqueous solutions, they adsorb and keep the dissolved materials (Akkaya et al. 2009; Rashidzadeh et al. 2015; Noein et al. 2017; Modarresi-Saryazdi et al. 2018). This water absorption is possible due to hydrophilic groups that establish hydrogen bonding with water (Abdollahi et al. 2016; Fallahi-Samberan et al. 2019) or electrostatic interactions (Lu et al. 2015a, b; Abdollahi et al. 2018). The water absorption process can be fine-tuned for diverse applications such as removal of dyes and heavy metals from wastewaters (Lu et al. 2015a; Li et al. 2011; Kasgöz and Durmus 2008) and drug delivery systems (Yassin et al. 2015; Fallahi-Sambaran et al. 2018; Dehghani et al. 2019). This is achieved by controlling the amount of water uptake through changes in pH or temperature of the medium (Banaei and Salami-Kalajahi 2015; Hernandez-Martínez et al. 2018; Nikravan et al. 2018a, b). Special structure of hydrogels allows the diffusion of solutes into the network. Also, hydrogel possesses a lots of ionic or nonionic functional groups those can absorb or trap ionic dyes from wastewater (Liu et al. 2010).

Incorporating natural materials into hydrogels affects their swelling behavior and adsorption capacity. Carbohydrate polymers are more attractive than synthetic ones due to their unique properties such as good hydrophilicity, biodegradation and biocompatibility (Sharma et al. 2015; Vakili et al. 2014; Abdeen and Mohammad 2014; Golshan et al. 2017). Cellulose is the most available and naturally renewable polysaccharide with unique physical and mechanical properties (Varaprasad and Jayaramudu 2017; Zhou et al. 2011). It contains hydroxyl groups those can help adsorption of water and dye molecules (Suo et al. 2007; Ibrahim et al. 2007). Many cellulose derivatives such as nanocrystalline cellulose (NCC) (Yang et al. 2012), carboxymethyl cellulose (CMC) (Nath and Dolui 2018) and hydroxyethyl cellulose (HEC) (Wang et al. 2010) have been used for synthesizing hydrogels adsorbents with acrylic acid. Zhu et al. reported the preparation of palygorskite/poly(acrylic acid) (PAA) nanocomposite hydrogels for adsorbing cationic basic dye (Zhou et al. 2011). Liu et al. reported β-cyclodextrin/PAA-grafted graphene oxide (β-CD/PAA/GO) hydrogels for removal of cationic dye, methylene blue (MB), from aqueous solution through electrostatic attraction, host–guest supramolecular interactions, and π-π conjugation interactions (Liu et al. 2014). Solpan et al. prepared poly(acrylamide-co-acrylic acid) poly(AAm-co-AA) hydrogel by irradiating with gamma radiation and applied it for the adsorption of cationic dyes such as Safranine-O (SO) and Magenta (M) (Şolpan et al. 2003). Singh et al. synthesized a novel biodegradable multifunctional superabsorbent hydrogels (SAHs) for adsorption of crystal violet (CV) and MB (Singh and Singhal 2015). Qi et al. fabricated adsorbents by cross-linking co-polymerization of acrylamide (AAm) and diallyldimethylammonium chloride (DADMAC) in the presence of salecan and used them for efficient removal of dyes and heavy metal ions (Qi et al. 2018). Zendehdel et al. prepared a semi-interpenetrating hydrogel composed of poly(AAm-co-AA) and poly(vinyl alcohol). Then, the adsorption ability of the hydrogel for removal of MB from aqueous solutions was investigated (Zendehdel et al. 2010). Zarezadeh-Mehrizi et al. used the octavinyl polyhedral oligomeric silsesquioxane nanoparticle (OVS) as a crosslinker for the PAA/OVS hydrogel with 3D crosslinking network via radical polymerization. The adsorption properties of the obtained nanocomposite hydrogels were tested for MB as a model cationic dye (Zarezadeh-Mehrizi et al. 2017).

In this study, crosslinked PAA/NCC nanocomposite hydrogels with different amounts of NCC are prepared by in situ free radical polymerization in aqueous media and used in removal of MB. To investigate the effect of surface modification on swelling behavior of nanocomposite hydrogels and dye removal capacity, NCC is used in pristine state and also, modified with (3-amino)propyltriethoxysilane (APTES) (NCC-APTES) and hexadecyltrimethoxysilane (NCC-HDTMS). All nanocomposite hydrogels are used as MB absorbent in different pH values because MB is the most commonly used cationic dye in industries and its effective removal from wastewater is very important. The high surface area and surface functional groups of nanoparticles strongly suggest affecting the adsorption capacity of hydrogels. Finally, the adsorption kinetics of MB by different nanocomposite hydrogels are studied using different adsorption kinetics models.

Experimental methods

Modification of NCC by APTES and HDTMS

For surface modification, NCC (0.7 g) was dispersed in ethanol (70 mL) ultrasonically to reach a homogeneous dispersion at 60 °C. After purging nitrogen, silane coupling agent (2.8 mL) (12.0 mmol APTES and 7.2 mmol HDTMS) was added dropwise during 2 h and the reaction was continued for 3 h. For purification, the obtained suspension was centrifuged (10,000 rpm) and washed with a solution of water/ethanol three times. The final product was then freeze-dried overnight (Zhang et al. 2012).

Synthesis of PAA/(modified) NCC nanocomposite hydrogels

The hydrogels were prepared by free radical polymerization of AA in NCC or modified NCC suspension using KPS as initiator. The nanocomposite hydrogels were synthesized by varying the amount of (modified) NCC from 1.0 to 3.0 wt% (with respect to the amount of monomer) at constant amount of crosslinker (N,N’-methylenebis(acrylamide), MBA) and initiator (KPS) Typically, a mixture of water (10 mL), AA (2.36 mL, 34.4 mmol), and (modified) NCC was prepared and the mixture was homogenized to ensure the NCC was suspended uniformly. Then, MBA (0.26 g, 1.7 mmol) in 5 mL water was added to the solution and heated in an oil bath at 60 °C. After replacing atmosphere by nitrogen, KPS solution (0.06 g, 0.2 mmol) was added and reaction was performed for 24 h. Final product was washed with distilled water for 3 days and the sample was dried in vacuum at 50 °C. For NCC-HDTMS containing hydrogels, ethanol was used as solvent. Table 1 shows the sample coding for modification and polymerization steps.

Table 1 Sample coding for modification and polymerization steps
Full size table

Evaluation of swelling behavior

The swelling behavior of nanocomposite hydrogels was evaluated by immersing dried hydrogels in aqueous media at different pH values. Then, the sample was taken out at certain time intervals, blotted with filter paper to remove water on the surface and weighted. The swelling ratio (SR) for each sample was calculated via Eq. (1) (Anirudhan and Rejeena 2012):

$$SR = \frac{{W_{s} - W_{d} }}{{W_{d} }}$$
(1)

where WS and Wd are the weights in swollen and dry states respectively.

Adsorption studies

All adsorption experiments were carried out at ambient temperature. Aqueous solutions of MB at two pH values (1 and 8) were used to investigate the adsorption kinetics and isotherms. Hydrogels (0.05 g) was immersed in MB solution (100 mL, 5 mg/L) and sampling was performed at pre-determined time intervals by taking out a 1 mL aqueous solution and replacing fresh water. The concentration of MB was measured by a UV–Vis-NIR spectrophotometer at 662.5 nm for pH = 1 and 665 nm for pH = 8. The adsorption capacities of hydrogels (qe, mg dye/g hydrogel) were calculated via Eq. (2) (Akkaya et al. 2009).

$$q_{e} = \frac{{(C_{i} - C_{e} )V}}{m}$$
(2)

where Ci and Ce are the concentrations of MB in the initial solution and the aqueous phase after adsorption respectively, V is the volume of the dye solution and m is the weight of hydrogel.

Results and discussion

In this study, we have prepared a series of PAA/(modified) NCC nanocomposite hydrogels containing different amounts of (modified) NCCs using free radical polymerization in the presence of MBA as cross-linker. At first, we modified NCC with APTES and HDTMS. Then, the free radical polymerization was performed to produce nanocomposite hydrogels those were used for adsorption of MB dye. The preparation route of nanocomposite hydrogels and adsorption behavior are shown in Scheme 1.

Scheme 1
scheme 1

Fabrication process of nanocomposite hydrogels those are used in dye absorption

Full size image

Preparation of NCC and its modification

To confirm the success of the preparation and modification of NCC, various analyses were performed. As observed by FE-SEM images (Figure S1), MCC have a length of 10–50 μm and a diameter of 10–25 μm. These large sized microcrystals are made from microfibers with strong hydrogen bonding. MCC was converted to NCC through acid hydrolysis. NCC is rod-like with length of 100–200 nm and diameter10–20 nm. NCC has a coagulated and aggregated morphology due to its high surface area and strong hydrogen bonding (Habibi 2014). Figure S2 shows the FT-IR spectra of MCC, NCC, NCC-APTES and NCC-HDTMS. According to the results, due to the intracellular hydrogen bonding of cellulose, the broad peak at 3200–3550 cm−1 is assigned to O–H groups (Haqani et al. 2017) and C–H stretching vibrations are observed at 2850–2950 cm−1 (Panahian et al. 2014). Peak at 1645 cm−1 originates from –OH stretching of water molecules absorbed to the structure of nanoparticles (Mazlita et al. 2016) and peaks appearing from 1310 to 1370 cm−1 are due to the bending vibration of the C–H and C–O groups in the polysaccharide rings in cellulose (Kargarzadeh et al. 2012). Also, peak at 1050–1110 cm−1 is due to the bending vibration of the C–O–C (Yang 2017). The characteristic peak of β-glucosic bond between the glucose rings and the cellulose chains are observed at 896–900 cm−1 (Yang 2017; Shankar and Rhim 2016). After modification of NCC with APTES, new peaks appeared at 3320 and 3290 cm−1 related to stretching and at 1564 cm−1 due to bending vibrations of –NH2 groups (Gao et al. 2009; Sharifzadeh et al. 2016). The typical absorption peaks of the Si–O–Si bonds of the siloxane compounds were overlapped by the cellulose bands due to C–O bending modes (Gao et al. 2009). In NCC-HDTMS, observed peaks at 2925 and 2854 cm−1 represent the symmetric and asymmetric stretching vibrations of CH2 groups (Amirshaqaqi et al. 2014). This indicates that the long-chain alkyl groups (–C16H33) of HDTMS were successfully grafted on NCC surface (Gao et al. 2009; Luo et al. 2017). The XRD patterns of MCC, NCC, NCC-APTES and NCC-HDTMS are shown in Figure S2. NCC displays four diffraction peaks at 15.1 (110), 16.5 (110), 22.7 (200), and 34.6° (004) those are typical peaks of cellulose (Kumar et al. 2014). Peaks at 15.1 and 16.5° show the cellulose type I structure. Crystallinity index (Ic) of samples was calculated according to Segal peak height method via Eq. (3) as follows:

$$I_{c} (\% ) = \frac{{I_{(200)} - I_{am} }}{{I_{200} }} \times 100$$
(3)

where I(200) represents combined crystalline and amorphous parts of cellulose (peak intensity at around 22.7°) and Iam represents the amorphous part of cellulose (peak intensity at around 18°) (Segal et al. 1959; Ahvenainen et al. 2016). According to the results, crystallinity of MCC, NCC, NCC-APTES and NCC-HDTMS was calculated 79.3, 82.7, 80.9 and 80.7% respectively. Decreasing the NCC crystallinity index after modification is associated to the APTES and HDTMS amorphous parts. However, XRD peaks of modified samples showed that the structure of NCC nanoparticles is retained and different processes including modification and acid washing did not disturb the crystalline structure. The size of crystallites for different samples was also calculated according to Scherrer Equation (Bhatia et al. 2017; Pirayesh et al. 2018):

$$D = \frac{K\lambda }{B\cos \theta }$$
(4)

where θ is the diffraction angle, K is the Scherrer constant, λ is the X-ray wavelength and β is the peak width at half of maximum intensity in radians. The crystallite size for MCC, NCC, NCC-APTES and NCC-HDTMS were 3.9, 4.1, 4.0, and 3.6 nm respectively. Higher crystallite size of NCC compared to MCC caused by degradation of smaller crystallites and growth of incomplete crystallites during the MCC acid hydrolysis process (Maiti et al. 2013). The NCC surface modification led to the dissociation of the hydrogen bonds in its structure which prevented aggregation of NCCs; thereby, the size of crystallites was decreased after surface modification.

According to TGA results (Fig. 1), weight loss of 79.1% in NCC is due to the loss of water in glucose units in the main chain, the break of the main molecular backbone and the failure of the C–O and C–C bands. The small particle size and high effective surface area of NCC led to its degradation at low temperatures. Also, amorphous chains on the surface of NCC degraded at lower temperatures and reduced its thermal stability. Degradation of NCC occurred at one step and there is no degradation step at temperatures between 150 and 200 °C which is due to the absence of the sulfate groups in NCC (Cha et al. 2012). So, repeated centrifugation and ultrasonic processes and adding NaOH to the solution resulted in sulfate free NCC product. After modification with APTES and HTDMS, weight loss reduced to 69.6 and 72.7% respectively. The thermal stability of nanoparticles was improved due to the crosslinking of the APTES and HDTMS via silanol groups on the surface of the nanoparticles.

Fig. 1
figure 1

TGA thermograms and DLS results of MCC, NCC, NCC-APTES and NCC-HDTMA nanoparticles

Full size image

DLS was performed to investigate the changes in size of nanoparticles after each step of modifications. To this end, a 1 mg/mL solution of MCC, NCC, NCC-APTES and NCC-HDTMS was analyzed at room temperature as results are depicted in Fig. 1. Z-average particle size of MCC, NCC, NCC-APTES and NCC-HDTMS was reported 503.6, 304.7, 45.7 and 121 nm respectively. It was observed that NCC has a smaller particle size than MCC due to the failure of glucose bands during acid hydrolysis and physical reduction of the dimensions (Nilsson 2017). The smaller size of modified samples indicated the breakdown of hydrogen bonding in NCC structure. In other words, surface modification of the nanoparticles prevented the accumulation of NCCs resulting in a decrease in particle size. It was also noted that NCC-HDTMS samples showed larger particle size than NCC-APTES due to the dispersion forces in NCC-HDTMS those acted as a driving force for aggregation. On the other hand, a narrow size distribution was observed in all samples. The PDI values obtained for MCC, NCC, NCC-APTES and NCC-HDTMS were 0.02, 0.07, 0.01 and 0.007 respectively. Also, the Zeta potential for MCC, NCC, NCC-APTES and NCC-HDTMS were calculated to be − 16.4, − 19.6, + 2.2, and − 2.1 respectively (Figure S2). The higher charge on NCC surface than MCC is due to acid hydrolysis process and the increase of the surface which led to an increase in OH groups on the surface and thus, increasing negative charge absorption rate on the surface (Melo et al. 2018). On the other hand, after modification of NCC with HDTMS, the charge on the surface was reduced, which is explained by replacing OH groups with carbon chains. In NCC-APTES, the zeta potential showed a positive amount that is related to the –NH2 groups of the APTES those can absorb protons in aqueous medium (Yamada et al. 2006).

As a conclusion, surface modification of NCC was performed successfully using hydroxyl groups on its surface. Controlling the hydrophilicity and hydrophobicity of NCC is very important to achieve the best compatibility between NCC and the polymer matrix. In this regard, we used two modification agents (APTES and HDTMS) to produce hydrophilic and hydrophobic NCCs those were incorporated in PAA matrix to fabricate nanocomposite hydrogels.

Fabrication and characterization of nanocomposite hydrogels

To investigate the properties of nanocomposite hydrogels, different analyses were utilized. The thermal degradation behavior of the nanocomposite hydrogels was studied in the temperature range of 100–700 °C under nitrogen atmosphere as results are presented in Fig. 2 and summarized in Table S1. PAA weight loss occurred in two stages. The first decomposition of PAA occurred with release of H2O, CH4 and monomer (acrylic acid) in the range of 150–330 °C and the second weight loss occurred between 300 and 600 °C due to the separation of the short chain of acrylic acid from the main chain (degradation of the backbone). It is clearly showed that incorporation of NCC into PAA hydrogel and interaction of NCC with PAA chains enhanced the thermal stability of the hydrogel even insignificant. This may originate from low thermal stability of NCC. The Td,max of HN0 was about 349 °C and shifted to higher temperatures by increasing NCC content. Similar trends were observed for HA and HH nanocomposite hydrogels where increasing the amount of nanoparticles from 1 to 3 wt% led to better thermal stability. On the other hand, it was observed that the HH hydrogels had a higher degradation temperature than HN and HA due to modification with HDTMS.

Fig. 2
figure 2

TGA thermograms of nanocomposite hydrogels prepared with different amounts of NCC, NCC-APTES and NCC-HDTMS

Full size image

XRD diffractograms for different nanocomposite hydrogels are shown in Figure S3. XRD patterns for hydrogels were found to be dramatically affected by the presence of PAA, where only a single broad peak ranging from 10° to 30°, characteristic of the amorphous domain, was observed. This observation could be explained by the small content of nanoparticles in the hydrogel nanocomposites (Hou et al. 2017) and appropriate dispersion of nanoparticles in polymeric matrix (Moqadam et al. 2018). FE-SEM images of HN, HA, and HH nanocomposite hydrogels are presented in Fig. 3. All nanoparticles are dispersed in matrix in aggregated form. On the other hand, the surface of nanocomposites is rough and has many undulations and folds. Due to the uneven surface morphology, the surface area of these hydrogels has increased and water molecules have penetrated the structure of hydrogels (Luo et al. 2018). Moreover, the increase of NCC values leads to the filling the hydrogels structure that may result in lower water diffusion to the structure of hydrogels (Bashir et al. 2017). However, in HH nanocomposites the agglomerations are bigger than other nanocomposite hydrogels. This can be ascribed to hydrophobic surface of NCC-HDTMS nanoparticles where dispersion forces are the main driving force to make bigger agglomeration.

Fig. 3
figure 3

FE-SEM images of NCC/PAA, NCC-APTES/PAA, and NCC-HDTMS/PAA nanocomposite hydrogels

Full size image

Swelling behavior of nanocomposite hydrogels

The swelling ratio was monitored for different nanocomposite hydrogels via gravimetry approach in pre-determined time intervals at different pH values as results are illustrated in Figs. 4 and S3–S5. Figure 4 shows the swelling ratio for HN2, HA2 and HH2 nanocomposite hydrogels where all samples contain 2 wt% nanoparticles. It was found that at all pH values, HA2 was the most swollen sample and HH2 was the least swollen one. This is likely due to presence of NH2 groups in HA samples which resulted in more water absorption. However, NCC-HDTMS had hydrophobic surface which resulted in less water absorption in nanocomposite hydrogels. Moreover, as pH value increased, deprotonation of COOH functional groups of PAA were favored and the electrostatic repulsions between the COO− anions caused the expansion of the hydrogel network and thus, swelling ratio (Kurdtabar et al. 2018; Nikravan et al. 2018a, b).

Fig. 4
figure 4

Swelling ratio of HN2, HA2 and HH2 nanocomposite hydrogels at different pH values

Full size image

Figure S4 shows the effect of NCC content on the swelling ratio at different pH values. At first, higher amount of NCC led to higher swelling ratio due to hydrophilicity of cellulose nanocrystals (Lim and Ahmad 2014). However, at pH values equal to or higher than pKa value of PAA [pKa = 4.25 (Lim et al. 2015)], carboxyl groups are partially or fully deprotonated which results in higher water absorption. In this state, strong hydrogen bonding between carboxyls of PAA and hydroxyls of NCC resulted in the more difficult penetration of water molecules into the structure. In the case of HA hydrogel nanocomposites (Figure S5), adding NCCs resulted in higher swelling ratio compared to neat hydrogel due to hydrophilic nature of NCC-APTES nanoparticles. However, adding more NCC-APTES nanoparticles resulted in decreasing swelling ratio. This can be attributed to the filling of the hydrogel network space with nanoparticles, which resulted in a more rigid hydrogel structure and subsequently, difficult water penetration into 3D structure of hydrogels (Lim et al. 2017). As depicted in Figure S6, an increase in the amount of NCC-HDTMS nanoparticles resulted in lower swelling ratio even compared to neat hydrogel. This is attributed to hydrophobic surface of nanoparticles due to modification with HDTMS. Also, higher amount of nanoparticles resulted in a rigid structure of hydrogel nanocomposite and filled-space of 3D hydrogel network. This favored difficult water penetration and led to less swelling ratio.

Removal of MB by nanocomposite hydrogels

Pollution of surface and ground water with synthetic dyes is a significant environmental problem and is a threat to human life (Gupta et al. 2013). MB as an extensively-used dye causes different eye and skin injuries (Ghosh and Bhattacharyya 2002). However, because of its importance in industry, the adsorption capacity of MB by hydrogel nanocomposites is subject of work. Adsorption properties of the hydrogels were evaluated by means of UV–Vis. absorbance at 662.5 or 665 nm depending on pH of medium. Generally, removal of dyes by hydrogels are affected by (1) the surface charge of the adsorbent, (2) the degree of ionization of the adsorptive molecule and (3) extent of dissociation of functional groups on the active sites of the adsorbent (Crini 2006). The pH value of the dye solution is an important parameter in its adsorption by hydrogels. At higher pH values, carboxylic acid groups are ionized. So, electrostatic repulsions between COO− groups leads to expanded structure of the hydrogels and increasing the interactions between negatively-charged groups in AA and positively-charged MB dye. Figure 5 shows the effect of pH value of the MB solution on the adsorption capacities of nanocomposite hydrogels containing 2 wt% of nanoparticles. It was found that as pH increased, the adsorption capacities increased for all samples. This can be explained by electrostatic interactions between MB and hydrogel structure. On the other hand, the –COOH groups dissociated to form COO− at higher pH values. Thus, carboxyl anions dominated the electrostatic repulsions between ionized groups leading to expansion of the hydrogels network and increasing adsorption capacity. Moreover, more deprotonated groups dominated interactions between cationic MB and hydrogel that resulted in higher adsorption capacity (Bulut and Karaer 2015). Surface modification of NCC also affected the adsorption capacity as shown in Fig. 5. According to the results, there is no significant difference between adsorption capacities by different samples at pH = 1. This originates from low swelling ratio of hydrogels that did not let MB to diffuse into 3D network. However, at pH = 8, HN2 nanocomposite hydrogel had the highest adsorption capacity whereas HA2 had the lowest one. This can be explained by presence of –NH2 groups on the surface of NCC-APTES that can prevent cationic dye adsorption via electrostatic repulsions.

Fig. 5
figure 5

Adsorption of MB versus time for HN2, HA2 and HH2 nanocomposite hydrogels at different pH values (n = 3, mean ± SD)

Full size image

Figure 6 shows that by incorporation and increasing the amount of all types of nanoparticles the adsorption capacity decreased at pH values of 1 and 8. This can be explained by filling the hydrogel structure by nanoparticles which did not permit to diffusion of dye from solution to 3D structure. However, by varying the pH from 1 to 8, the adsorption capacity increased for all samples. For example, HN2 nanocomposite hydrogel showed equilibrium point at 2.6 mg/g at pH = 1 whereas at pH = 8, the equilibrium increased to 38 mg/g. This showed that the alkali media is better for MB adsorption because COOH groups are deprotonated and more interaction between the cationic dye and COO− leads to increasing adsorption.

Fig. 6
figure 6

Adsorption of MB versus time for HN, HA and HH nanocomposite hydrogels containing different amounts of nanoparticles at different pH values

Full size image

Adsorption kinetics

Prediction of adsorption kinetics is necessary for good understanding the adsorption mechanism. To investigate the adsorption mechanism of the nanocomposite hydrogels for MB dye in aqueous solutions, the adsorption kinetics was analyzed. Three kinetic models were applied to fit the experimental data including pseudo-first-order (Eq. 5), pseudo-second-order (Eq. 6) and intra-particle diffusion (Eq. 7) models (Bulut and Karaer 2015; Chen and Wang 2009).

$$\ln \left( {q_{e} - q_{t} } \right) = \ln q_{e} - k_{1} t$$
(5)
$$\frac{t}{{q_{t} }} = \frac{1}{{q_{e}^{2} k_{2} }} + \frac{t}{{q_{e} }}$$
(6)
$$q_{t} = k_{id} t^{1/2} + C$$
(7)

where qe and qt (mg/g) represent the amount of dye adsorbed on nanocomposite hydrogels at equilibrium and any time respectively and k1 (min−1 or h−1) is rate constant. qe and k1 can be obtained by the intercept and slope of ln(qe-qt) versus t (h) (Figure S7). k2 (g/(mg h)) is the pseudo second-order rate constant where qe and k2 can be calculated from the slope and intercept of t/qt against t (h) (Figure S8). kid (mg/(g h1/2)) is the intra-particle diffusion rate parameter where plot of qt against t1/2 should be a straight line with a slope of kid and intercept of C (Figure S9). The calculated qe, k1, k2, kid, C and the corresponding linear regression correlation coefficient (R2) are summarized in Tables 2, 3 and 4. It can be concluded that both pseudo-first-order model and pseudo-second-order model have good correlation coefficients for MB dye, but the pseudo-second-order model have a better fitting data. Also, difference of qe values between the experiment and calculated values at pseudo second-order model is slightly higher than pseudo-first-order model.

Table 2 Adsorption kinetic parameters of pseudo-first-order model for MB adsorption onto hydrogels
Full size table
Table 3 Adsorption kinetic parameters of pseudo-second-order model for MB adsorption onto hydrogels
Full size table
Table 4 Adsorption kinetic parameters of intra-particle diffusion model for MB adsorption onto hydrogels
Full size table

Intra-particle diffusion model showed that the adsorption mechanism for dye had two separate phases due to the appearance of the two sections with variant linear slopes; the first sharper segment is due to the transportation of dye molecules from the bulk solution to the external surface of the hydrogels by diffusion through the boundary layer (film diffusion), and the second portion comes from the equilibrium stage as the dye concentration was very low. It was found that the correlation coefficient for the intra-particle diffusion model is lower than the pseudo-first-order and the pseudo-second-order models. This indicates that the intra-particle diffusion model is not appropriate for adsorption of MB onto the nanocomposite hydrogels.

Conclusions

In this study, we prepared a series of PAA/(modified) NCC nanocomposite hydrogels containing different amounts of (modified) NCCs using free radical polymerization in the presence of MBA as cross-linker. At first, we modified NCC with APTES and HDTMS. Then, the free radical polymerization was performed to produce nanocomposite hydrogels those were used for adsorption of MB dye. HA nanocomposites showed the most swollen state whereas HH ones had the least SR due to existence of hydrophilic NH2 groups in HA samples and hydrophobic surface of nanoparticles in HH nanocomposites. Moreover, higher pH values favored deprotonation of COOH functional groups of PAA and the electrostatic repulsions between the COO− anions which caused higher SR. MB adsorption of the hydrogels showed that removal of MB were affected by the surface nature of nanoparticles, pH value of medium and amount of nanoparticles. It was found that as pH increased, the adsorption capacities increased for all samples because more deprotonated groups dominated interactions between cationic MB and hydrogel. However, HN nanocomposite hydrogels had the highest adsorption capacity whereas HA samples had the lowest one due to presence of –NH2 groups on the surface of NCC-APTES that can prevent cationic dye adsorption via electrostatic repulsions.