Synthesis of a nanocomposite with holocellulose extracted from barley straw and montmorillonite, and optimization of the removal of methylene blue dye using the synthesized adsorbent

In this study, holocellulose was extracted from milled barley straw with different mesh sizes using Laccase enzyme. After extraction, a dual composite was made using montmorillonite clay to remove methylene blue dye from synthetic effluent. Results of different analysis methods including scanning electron microscope, Fourier transform infrared spectroscopy, and BET revealed that prepared nanocomposite presented desired specifications, and for smaller mesh sizes, derived holocellulose had higher quality due to high specific surface area. Response surface methodology was employed to reduce the number of experiments for methylene blue adsorption experiments and to achieve an empirical model for prediction of adsorption efficiency at different operating conditions. The effect of different factors including solution temperature, pH, initial dye concentration, and mesh size of milled barley straw on dye adsorption performance by prepared composite was examined. Maximum removal efficiency was obtained about 95% at temperature of 32 °C, pH of 8, initial dye concentration of 4 mg L−1, and mesh size of 70. Also, isotherm studies were performed on experimental data using Langmuir, Freundlich, Temkin, and Dubinin–Radushkevich isotherm models and results revealed that the adsorption process follows Langmuir model with maximum predicted adsorption capacity of 159 mg g−1, which implies monolayer adsorption. Moreover, thermodynamic study revealed that adsorption of methylene blue is endothermic and spontaneous while enthalpy and Gibbs free energy of adsorption are positive and negative, respectively. Finally, adsorption kinetic study determined that the pseudo-second-order kinetics model with correlation coefficient of about 1 best fitted the experimental results which is the characteristic of chemisorption process.


Introduction
One of the principal problems of developing countries is contamination of water resources because of the development of associated industries.Toxic and hazardous materials such as polymers, dyes, organic pollutants, and heavy metals can be detected in wastewater of various industries such as textile, paper, painting, leather, plastics, cosmetics, and food, most of which are carcinogens (Katheresan et al. 2018;Malyan et al. 2019;Monash & Pugazhenthi 2009).About 10-15% of industrial dyes are left untouched in the dyeing process and enter the environment, specifically water resources.Treatment and recovery of municipal and industrial wastewater to prevent environmental contamination and its reuse have become increasingly prominent in recent decades.Different methods have been proposed and used to remove organic dyes from industrial effluents including chemical oxidation, flocculation, photolysis, adsorption, reverse osmosis, etc. (Katheresan et al. 2018;Bhatia et al. 2017).Most of these methods usually are not economic.Meanwhile, adsorption with bio-based adsorbents is growing as a high-performance and low-cost approach (Yagub et al. 2014;Katheresan et al. 2018).Todays, various types 243 Page 2 of 20 of food and agricultural by-products such as barley straw, wheat straw, rice paddy, chitosan, banana peel, sawdust, sugarcane and coconut stalks, and orange peel are used as enzymatic adsorbents to remove contaminants from various water resources (Yagub et al. 2014;Mostafa et al. 2021;Ahmed & Hameed 2018;Akhayere & Kavaz 2021;Husseien et al. 2007;Laribi et al. 2020).
Cellulose is one of the most abundant natural bio-based polymers on earth.Despite its relatively simple chemistry, the physical structure and morphology of cellulose in plants are complex and heterogeneous (Ludueña et al. 2013).Due to easy maintenance, availability, high specific surface area, suitable special mechanical properties, cost-effective production, safer application, biodegradability and renewability, thermal stability, lower energy consumption for production, and many other advantages, it is a very good alternative to petroleum and plastic materials (Dufresne 2013;Phanthong et al. 2018).Hemicellulose fibers have attracted the attention of many researchers in recent years due to their unique qualities.These fibers are available in many forms, such as films and gels and can be mixed with other materials such as polymers to make composites.They are also used in many fields such as bio-nanocomposites, particularly in the automotive and aerospace industries, packaging industry, production of new filters, gas-impermeable films, and hydrogels.(Siqueira et al. 2010;John & Thomas 2008;Lucenius et al. 2019;Chen et al. 2021).Holocellulose nanocrystal fibers by definition are composed of cellulose and partially hemicellulose fibers while lignin and other extractives are removed (Rabemanolontsoa & Saka 2012;Jiang & Hsieh 2015;Chen et al. 2021).Holocellulose extraction is lowcost compared to cellulose extraction because it needs less energy and chemicals consumption and with recyclability property (Yang & Berglund 2020;Laribi et al. 2020).It combines advantages of cellulose and hemicellulose.It has many applications such as preparation of nano-papers (Galland et al. 2015;Park et al. 2017), composite films (Šimkovic et al. 2017;Li et al. 2017b), thermal insulation aerogels for freeze drying (Huang et al. 2020), and printing inks (Chen et al. 2021).Properties, benefits, and applications of holocellulose fibers are completely described by Yang and Berglund (2020).Despite cellulose nanofibers which are hydrophilic, holocellulose nanofibers are amphiphilic which makes them suitable for use in both polar and non-polar environments (Jiang & Hsieh 2015).
Nanocomposite materials have been extensively studied and researched in recent years.They have found many applications in different fields such as water and wastewater treatment (Etemadi et al. 2021;Beyene & Ambaye 2019;Hnamte & Pulikkal 2022;Adeola & Nomngongo 2022), gas storage (Djeridi et al. 2019;Rezayeenik et al. 2022;Srinivasan et al. 2020), catalysts for different chemical reactions (Zinatloo-Ajabshir et al. 2019;Gerasimov 2022;Liu et al. 2022;Stevanović et al. 2022;Cyganowski & Dzimitrowicz 2022), medicine and treatment (Ramezani & Ripin 2023;Qamar 2021;Huang et al. 2022;Nasef et al. 2023), thermal energy storage (Li et al. 2017a(Li et al. , 2023;;Haghighi et al. 2020), and environmental remediation (Mohamed 2022;Singh & Ambika 2018;Goci et al. 2023).At least one of the phases of a nanocomposite has one or more dimensions less than 100 nm and nano-scaled particles are distributed within the dominant phase to increase robustness, resistance, electrical conductivity, and magnetic properties (Camargo et al. 2009).To achieve better properties in nanocomposites, the size distribution of nanoparticles and their dispersion within the matrix of the composite should be controlled.Cellulosebased nanofibers are used for reinforcement of composites materials.These fibers provide robustness and rigidity, and their intrinsic properties determine characteristics of composites (Rouquerol et al. 2013;Ardanuy et al. 2015;Kalia et al. 2011). Hosseini et al. (2022) utilized ternary nanocomposite of carboxymethyl cellulose grafted by polyacrylic acid and decorated with graphene oxide for the removal of methylene blue (MB) and showed that the prepared composite has maximum MB adsorption capacity of 138 mg g −1 after 250 min.Moreover, they conducted reusability tests and showed that 90% of nanocomposite capacity remained unchanged after 9 cycles.Munonde et al. (2023) investigated adsorption removal of MB using MnO 2 -reduced graphene oxide (rGO) nanocomposite adsorbent synthesized from rGO nanosheets and MnO 2 nanoparticles, assisted by sonication.The adsorption isotherm study showed that Langmuir isotherm model could fit experimental results more accurately with the maximum adsorption capacity of 156 mg g −1 .Also, kinetic study revealed that pseudo-second-order model could predict dynamic adsorption results which is the characteristic of chemisorption.They obtained maximum dye removal efficiency of about 98%.
Holocellulose can be extracted from various plant sources.Although wood is the most important source of holocellulose industrial production (Yang & Berglund 2020;Macfarlane et al. 1999), it has received less attention than other uses of wood (Chen et al. 2011a;Yang & Berglund 2021).Production of fibers from agricultural products such as flax, hemp, cotton, and especially agricultural by-products is of particular interest.Agricultural by-products include wheat and barley straw, rice bran, sugarcane stalks, fruit and vegetable pomace, etc. (Reddy & Yang 2005;Wang et al. 2022;Pennells et al. 2020).These non-wood materials contain less lignin than wood (Poletto 2018;Wang et al. 2022), so their decolorization process is less damaging to the fibers.Agricultural by-products are good candidates for use as adsorbents because they are abundant in nature, cheap and effective, and do not require special operations (Chen et al. 2011b;Mostafa et al. 2021;Ungureanu et al. 2023).Yao et al. (2015) developed extraction of hemicellulose by hot water using sodium hydroxide in order to decrease cellulose degradation and to achieve higher molecular weight of hemicellulose fibers.They utilized response surface method in order to optimize hemicellulose extraction process.Composition and purity of hemicellulose fibers were analyzed by high-performance liquid chromatography, FTIR, and SEM analysis methods and the results revealed that the solid residual silane extract was similar to commercial hemicellulose extract.Laribi et al. (2020) extracted holocellulose and cellulose from wheat straw and converted them to carboxymethyl holocellulose (CMH) and carboxymethyl cellulose (CMC) and studied their physicochemical properties using FTIR, XRD, TGA, and SEM analysis methods.Also, rheological study of the products showed that their aqueous solutions exhibited non-Newtonian thixotropic behavior.Wang et al. (2022) extracted holocellulose nanofibers from apple rhubarb pomaces using peracetic acid delignification and low energy blending.Diameter of extracted fibers ranged from 29.8 to 54 nm which was lower that diameter of obtained holocellulose fibers from wood (62.2 nm).Kwon et al. (2015) investigated adsorption of lead on holocellulose aerogels synthesized from wood fibers using hydroxide-urea solution.They studied different factors and showed that by increasing equilibrium concentration of lead and pH of the solution, the adsorption rate was increased.Therefore, they concluded that holocellulose aerogels can effectively be used for the removal of heavy metals, specially lead from water and wastewater.Wu et al. (2017) thermally bonded carboxymethyl chitosan-hemicellulose (CMCH) resins and examined them using FTIR, TGA, and SEM.Then, they studied adsorption characteristics of CMCH as adsorbent of metal ions of Cd (II), Ni (II), Cu (II), Hg (II), Mn (VII), and Cr (VI).Different factors affecting adsorption performance such as pH, temperature, contact time, and initial concentration of metal ions were investigated.Their results showed that the ultimate adsorption process follows quasi-second-order kinetic model and the equilibrium isotherm follows Langmuir model.Their experiments revealed that CMCH can significantly be used as a conceivably recyclable and effective adsorbent for the extraction and recovery of heavy metal ions from wastewater.Ahmed and Hameed (2018) utilized abundant wastes of barley straw as precursor to prepare biochar adsorbent through thermal pyrolysis.Pore characteristics, morphology, and attraction extent of the adsorbent toward salicylic acid were identified.The effects of batch adsorption variables, such as initial salicylic acid concentration, contact time, initial pH, and temperature, on biochar adsorption performance were tested.The salicylic acid/biochar system exhibited spontaneous, chemisorption, and endothermic natures.Considering its favorable pore characteristics and high attraction performance, barley straw biochar could be a suitable adsorbent to eliminate salicylic acid from wastewaters.Santawee et al. (2019) studied adsorption capacity of holocellulose and lignin extracted from coir pith for the removal of trimethylamine (fishy odor).Their results showed that holocellulose extracted from coir pith had higher removal percentage compared to commercial activated carbon.Akhayere et al. (2020) used barley husk waste as a precursor for the synthesis of nano-silica by following a low energy and sustainable method of acid reflux and heat treatment.Nano-silica was then used for the synthesis of magnetic nano-silica (M-NS) by adding magnetic solution of Fe 3 O 4 nanoparticles.Magnetic nano-silica particles were observed to have an average diameter of 162 nm and appeared to be hydrophobic, with a large surface area of ~ 120 m 2 g −1 and was used as an adsorbent for the removal of petroleum contaminants from water.Results showed that with 0.6 g of M-NS, high uptake efficiency of about 85% was achieved.Soldatkina and Yanar (2023) modified barley straw (BS-C) and corn stalks (CS-C) by citric acid for the purpose of dye pollutants removal from aqueous solutions.They used prepared adsorbent materials for the removal of MB and malachite green (MG) from wastewater.They examined effects of different parameters including pH, contact time, dye concentration, and adsorbent dosage on removal efficiency of cationic dyes and achieved maximum removal efficiency of 97% for MB removal with both adsorbents and 95% and 94% for MG removal with BS-C and CS-C adsorbents, respectively.
In recent years, montmorillonite has been used for the removal of different pollutants from wastewaters such as the removal of heavy metals (Alandis et al. 2019;Akpomie & Dawodu 2016;Sen Gupta & Bhattacharyya 2012;Praus et al. 2008;Naz & Chowdhury 2022), adsorption of textile dyes (Husseien et al. 2007;Cottet et al. 2014;Chen et al. 1999;Almeida et al. 2009;Nogueira et al. 2009), and natural and chemical organics (Yamamoto et al. 2017;Burgos et al. 2002).Also, Zhu et al. (2016) and Zango et al. (2022) reviewed different montmorillonite-based adsorbents for the removal of contaminants.Cottet et al. (2014) studied characteristics of montmorillonite clay modified by iron oxide for the adsorption and removal of MB in aqueous solutions.They reported maximum adsorption capacity of 71.12 mg g −1 at 333 K. Also, their results revealed that the adsorption process was endothermic and non-spontaneous and also pseudo-second-order kinetic model best fitted the experimental results.However, in another study, Almeida et al. (2009) reported that the adsorption of MB on montmorillonite was spontaneous yet endothermic.Their results also showed that maximum adsorption was achieved at pH of 11, temperature of 60 °C and adsorbent concentration of 3 g L −1 .Wang et al. (2020) synthesized montmorillonite supported-poly (acrylamideco-acrylic acid) hydrogel via polymerization and utilized the prepared adsorbent for the removal of MB from aqueous 243 Page 4 of 20 solutions.They reported maximum dye removal efficiency of about 97% at high concentration (200 mg L −1 ).Investigation of adsorption kinetics and isotherm revealed that pseudo-second-order kinetic together with Sips isotherm model could predict experimental results with high accuracy.
In this study, firstly barley straw was grinded into different mesh sizes, and holocellulose nanofibers were extracted from the particles.In the next step, montmorillonite nanoparticles were added to nanofibers in order to obtain a bio-nanocomposite for the removal of MB from aqueous solutions.Finally, different factors for the removal of MB, including solution pH, temperature, nanocomposite particle size distribution, and initial dye concentration, were optimized experimentally based on experimental design and response surface methods.To the best of our knowledge, it is the first time that contaminant removal by holocellulosemontmorillonite (HCMM) composite obtained from barley straw has been performed.Due to abundancy of barley straw and montmorillonite, this work can have economic as well as environmental benefits compared to other adsorbents.

Material and methods
Table 1 shows materials used in this study, their purity and manufacturer.Properties and sources of barley straw and montmorillonite clay was described separately at below.
Barley straw, as a native product of East Azerbaijan province in Iran, was used as the source of holocellulose.For this purpose, some barley straw was provided and ground into very fine particles and then meshed with 50, 100, and 150 mesh number sieves.The samples were dried in an oven at 40 °C for 24 h and then were dewaxed by ethanol and filtered.Filtered samples were dried again in an oven (Fan Azma Gostar, Iran) at 70 °C for 2 h.In the next step, 200 g of pure sample were poured in 1000 mL of distilled water and 2 g of Laccase enzyme was used to remove proteins.Filtered and dried sample was poured in 50 mL distilled water and then sonicated with ultrasound device (Elmasonic, Elma, Germany).Then, the pH of solution was set at 3 by adding hydrochloric acid into the solution and the sample was left at rest in order to precipitate and remove lignin.Finally, holocellulose nanofibers were extracted by filtering the solution and gathering them on a paper.Extracted holocellulose fibers were dried in an oven at 40 °C for 24 h.
The amount of holocellulose available in barely straw was evaluated using the following equation: where HC% is weight percent of holocellulose in barley straw, and m 1 and m 0 are weights of extracted holocellulose and dried barley straw samples, respectively.
Calculation of moisture content of samples was performed using the following equation: where W% is weight percent of moisture, and w 1 , w 2 , and w 3 are crucible weight, total weight of barley straw and crucible before drying, and total weight of dried barley straw and crucible, respectively.
Functional groups of extracted holocellulose were identified using a Fourier transform infrared (FTIR) spectrometer (Bruker Vector 22, Triad Scientific, USA).Complete description and analysis of the recorded spectra are presented in the Results section.
For preparation of HCMM composite, 7 wt% of sodium hydroxide and 24 wt% of urea were dissolved in distilled water, and temperature of the solution was decreased to 0 °C by placing it in an ice bath.Then, 5 wt% of holocellulose and 5 wt% of montmorillonite clay were added to the solution at that temperature.Finally, the prepared composite was placed in a 2 M hydrochloric acid solution and held in it for 24 h.The prepared composite was then washed by distilled water.
Specific surface area of the composite was calculated using BET analysis method which was performed by Quantachrome ChemBET 3000 device (Jiang & Hsieh 2015;Santawee et al. 2019).
Also, a sample of adsorbent was coated with a very thin layer of gold and then SEM images of the sample were taken using TESCAN MIRA3 FEG-SEM (Czech Republic).
Prepared composite was used for the removal of methylene blue (MB) from aqueous solutions.For this purpose, aqueous solutions with different concentrations of MB were prepared.Then, 0.1 g of HCMM composite was emerged in 100 mL of each solution for the adsorption and removal of MB.After 2 h, a sample from each solution was taken and concentration of remained MB in the solution was recorded using a spectrophotometer (Shimadzu, Japan) at maximum wavelength of 665 nm (Melgoza et al. 2009 The correlation between absorbance and concentration of MB was obtained from the above curve as the following equation: where A and C (ppm) are absorbance and concentration of MB in aqueous solution, respectively.
Percentage of MB removal, R%, and adsorption capacity of the composite, q e (mg g −1 ) were calculated using the fol- lowing equations: where V (L), is volume of the solution, m (g) is the mass of adsorbent, and C 0 and C e (mg L −1 ) are initial and equilibrium dye concentrations in the solution, respectively (Namjoufar et al. 2021;Senthil Kumar et al. 2014).
(3) C = 0.161A + 0.008 Experimental design and modeling using response surface method The experimental design and response surface methods (RSM) are helpful for reducing the number of required experiments and finding optimum conditions of a process when the desired response is affected by several variables and a theoretical model is hard to obtain.Usually, first-or second-order models with respect to input variables are used for empirical modeling of the response of a process (Eqs.( 6) and ( 7)) (Myers et al. 2016).
where y is output or response of the process, x i , i = 1, 2, … , k are input or independent variables, and is model error.
For calculation of unknown constants of the above models, several experiments should be conducted using experimental design method.There are many fractional or full factorial methods that can be used for this purpose.In this work, four input variables (factors) including barley straw mesh size, initial dye concentration in the solution, solution temperature, and solution pH were selected as input variables.One of the best second-order experimental design methods is central composite design (CCD) method (Myers (6)  Table 2 shows ranges of the input variables and their levels.Levels − 1, 0, 1 are defined for lower limit, middle value, and higher limit of the factors.Uptake amount of MB by the adsorbent after 120 min of contact time was considered as the response.

Adsorption isotherms
Four adsorption isotherm models including Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich (D-R) were used to identify the equilibrium mechanism of adsorption onto the synthesized composite.Linear forms of these models with respect to their coefficients are presented in the following equations (Namjoufar et al. 2021 where C e (mg L −1 ) is equilibrium concentration of dye on adsorbent surface in the solution, q e (mg g −1 ) is the amount of adsorbed dye at equilibrium, q m (mg g −1 ) is maximum adsorption capacity of the adsorbent calculated by Langmuir method, k L (L mg −1 ) is Langmuir constant, k F (mg g −1 ) is Freundlich constant, n is considered as the heterogeneity of the adsorbent surface, B = RT∕ b is a constant related to heat of adsorption, k T (L mg −1 ) shows equilibrium binding (8) is the universal gas constant, b is a parameter which is also related to heat of adsorption like B, q D and k D are D-R coefficients, and C s (mg L −1 ) is solubil- ity of MB in water.Energy of adsorption can be calculated using k D as (Pan et al. 2017; Togue Kamga 2018):

Thermodynamic study
Gibbs free energy of adsorption can be written as follows: where K C is distribution coefficient of adsorption and is defined as (Chowdhury et al. 2011): where C a (mg L −1 ) is equilibrium concentration of adsorbed material on the adsorbent and ads (kg m −3 ) is adsorbent density.For estimation of heat of adsorption, the Clausius-Clapeyron equation for liquid phase can be applied as follows (Builes et al. 2013): By applying the above equation on Temkin and D-R isotherm models, Eqs. ( 16) and ( 17) can be obtained for the calculation of heat of adsorption, respectively: where ΔH fus is the heat of fusion of MB.Equation ( 17) is obtained using the following equation (Salimi & Roosta 2019): where x MB , MB , T t , and ΔC P are soluble mole fraction of MB, activity coefficient of MB in liquid phase, fusion temperature, and solid-liquid heat capacity difference of MB, respectively.It was assumed that MB and ΔH fus are week functions of temperature and ΔC P is zero.According to the work of Salimi and Roosta (2019), solubility of MB at 32 °C was estimated as 54.33 g L −1 by linear regression and heat of fusion of MB was set to 26.22 kJ mol −1 .

Modeling of the kinetics of dye removal by the composite adsorbent
Kinetics of adsorption process can give valuable information about the rate of adsorption which is necessary for process design and control.In last part of the study, an aqueous solution containing 60 ppm of MB was prepared and kinetic experiments were performed.In certain time periods, a sample from the solution was taken and concentration of MB within the solution was obtained by spectrophotometry and then MB uptake amount was calculated using the following equation (Namjoufar et al. 2021): where q t (mg g −1 ) is the adsorbed dye uptake amount by the adsorbent, and C t (mg L −1 ) is concentration of dye in the solution at time t (min).Obtained experimental data were used for kinetic modeling.In this study, three different models including pseudo-first-order (linear driving force), pseudo-second-order, and intraparticle diffusion models were assessed and their parameters were obtained by fitting experimental data on them.Linear driving force (LDF) model can be written as follows (Naidu & Mathews 2021;Mahmoodi et al. 2012): where k 1 (min −1 ) is the rate constant of the model.This model states that the rate of adsorption is directly proportional to the number of free sites of the adsorbent (Mahmoodi et al. 2012).Integration of the above model and rearranging will result in the following equation (Simonin 2016): Therefore, if pseudo-first-order model holds for the provided data, plot of log q e − q t with respect to time should give a straight line.
The pseudo-second-order equation can be written as: where k 2 (min −1 ) is the reaction rate constant of pseudo-sec- ond-order model (Namjoufar et al. 2021).Integration of the above model with respect to time will give (Pan et al. 2017): which means that if this model holds, the diagrams of t q t with respect to t should give a straight line (Wang et al. 2016).Another kinetic model that was used in this study is intraparticle diffusion kinetic model, which can be written as (Mane et al. 2007;Pan et al. 2017;Cheng et al. 2020): where k 3 (mg g −1 min −0.5 ) is intraparticle diffusion rate con- stant.In order that this model to fit the experimental data, diagram of q t with respect to t 0.5 should give a straight line (Pan et al. 2017;Mane et al. 2007).The value of I gives an idea about the thickness of adsorption boundary layer.The larger the intercept, the thicker is the adsorption boundary layer (Mane et al. 2007).

Results of SEM analysis
SEM images of barley straw, extracted holocellulose, and prepared bio-nanocomposite, are shown in Fig. 2a-c.As can be seen, barley straw nanofibers are in the form of irregular aggregates.After extraction, the form of holocellulose is almost fibrous (Fig. 2b).It is also obvious from Fig. 2c that after preparation of the composite, the particles and aggregates of montmorillonite are distributed almost uniformly between holocellulose fibers.The aggregate particles are in cubic form with a size of about 500-900 nm in at least one dimension which confirms that a nanocomposite material has been synthesized.The difference in contrast of clay nanoparticles and holocellulose fibers is due to the difference in their molecular mass (Goldstein et al. 2018).

BET analysis results
As mentioned before, BET analysis was performed on HCMM composite in order to calculate average particle size, pore volume, and surface area of the adsorbent.According to the results shown in Table 3, high specific surface area was obtained and the composite can be classified as macroporous material due to relatively large average pore size (Somo et al. 2022).
Porosity of composite material can be obtained using the following equation (Hussaini & Dvorkin 2021) assuming homogenous spherical particles: (24) q t = k 3 t 0.5 + I where S (1/m), d p (m), and are specific surface area, aver- age particle size, and porosity of HCMM, respectively.Using the data from Table 3 and assuming that density of (25) S = 6 1 − d p non-porous HCMM is about 900 kg/m 3 by neglecting the effect of MM nanoparticles on the density of adsorbent (Sun 2005), porosity of HCMM was obtained about 0.7.Therefore, the density of porous HCMM can be estimated as:

Results of FTIR analysis
Results of FTIR analysis of extracted holocellulose for the three mesh sizes of barley straw are shown in Fig. 3a-c.In  the range of 3400-3450 cm −1 , wide and strong peaks are seen, which indicate the stretching vibrations of the O-H groups.The presence of these vibrations indicates extraction of cellulose and hemicellulose, due to the existence of many O-H groups in the structure of these compounds.Also, in the range of 2920-2930 cm −1 , a relatively strong peak is observed, which is due to the stretching vibrations of the C-H groups in the cellulose structure.Various peaks in the range of 1730-1750 cm −1 are evident in all three samples which are related to the vibrations of the C=O group in the hemicellulose structure.A relatively slight peak is also observed in the range of 1640-1650 cm −1 , which indicates O-H group related to water adsorption.Fine absorption peaks appear in the range of 1250 to 1450 cm −1 which correspond to the symmetric and asymmetric vibrations of C-H groups.Another relatively strong peak in the range of 1050 to 1150 cm −1 is related to the C-O group vibrations found in both cellulose and hemicellulose (Atykyan et al. 2020;Galletti et al. 2015).It is worth to note that by decreasing the particle sizes of barley straw (increasing mesh sizes), the intensities of peaks related to O-H and C=O increase sharply.These peaks indicate that extraction of hemicellulose increases with increasing mesh size.

Optimization of the removal of methylene blue adsorption by HCMM
Table 4 shows experimental design matrix and also experimental results of MB removal%.As it can be seen, minimum and maximum removal percentages are 77.24 and 92.14.
Based on the experimental results, a second-order model was fitted to find relations between effective factors and the MB removal amount by synthesized adsorbent using linear regression which is presented as the following equation.
where R is the estimated removal efficiency by the regres- sion model.Results of the analysis of variance of the above ( 27) Figure 4 depicts estimated values of MB removal efficiency calculated by the model compared to the experimental ones which indicates high accuracy of the model for prediction of experimental results within the domains of independent variables.

Effect of pH and temperature on removal efficiency
Figure 5a shows the effect of temperature and pH on MB removal efficiency.As it is evident, the rate of adsorption increases by increasing pH.Maximum adsorption is achieved within temperature range of 30 and 35 °C and pH of 8. Also, at acidic pH values, adsorption amount is decreased.As Österbeg and Claesson (2000) reported, cellulose fibers become compact at acidic pH values and thus available surface for adsorption decreases, while at pH value of 7.3 the fibers swell considerably and thus specific surface area increases.Also, at lower pH values, the composite  becomes positively charged and while MB is a cationic dye, the repulsion force between dye and composite reduces its adsorption.At alkaline pH values, the composite becomes negatively charged and attracts MB cationic dye, and therefore, by increasing pH, adsorption efficiency increases.For pH values above 9, it has been reported by (Mills et al. 2011) that hydrolysis decomposition of MB takes place and for that reason, experiments were conducted below this critical value.

Effect of temperature and initial MB concentration
Figure 5b illustrates the interactive effect of initial dye concentration and temperature on adsorption efficiency.
As can be seen, the extremum point is within temperature range of 30 and 35 °C and initial dye concentration of 4 mg/L.At temperatures lower than 30 °C, the rate of adsorption diminishes because adsorption of MB is endothermic which will be shown next when presenting results of adsorption isotherms.Also, at temperatures above 35 °C, dye adsorption uptake is reduced which can be deduced to deformation of holocellulose fibers and their thermal expansion.The optimum value for initial concentration of dye is the minimum value tested which is normal while higher concentrations of dye fill active sites of the composite more rapidly and reduce adsorption efficiency.Due to the elliptical nature of the diagram, it can be reasoned that the interaction between these two variables is significant on adsorption efficiency.

Effect of temperature and mesh size on the adsorption rate
Contours and optimum point for these parameters are shown in Fig. 5c.The extremum is at temperature of about 32 °C.In this case, the optimum value for mesh number, which is about 70, is not its lower limit which reflects interactive effect of temperature on this parameter.It could be due to the agglomeration of composite particles at higher temperatures and lower mesh sizes and reduction of available surface area for adsorption.Also, at higher temperatures, available surface area is decreased due to thermal expansion of fibers and filling of empty spaces.The optimum values of temperature and barley straw mesh size are not located at their assessed lower or higher limits which means that the selected domains for these variables are correct and contain optimum point.

Effect of initial dye concentration and mesh size of barley straw
Figure 5d shows combined effect of initial dye concentration and mesh size of barley straw, used for holocellulose extraction, on the rate of adsorption.As it can be seen, this diagram has a maximum value within the domains of these input variables which means that the assessed parameters have strong interactions with each other.Also, as it was predictable, the maximum adsorption rate was achieved at minimum value of initial concentration of dye, as higher initial dye concentration will decrease adsorption rate due to more rapid filling of active sites of composite.Also, maximum removal efficiency is obtained at smaller values of barley straw mesh size but not its lower limit, because although smaller mesh sizes result in higher amounts of hemicellulose in the extracted holocellulose and higher available surface area for adsorption, but at the same time, much higher specific surface area results in more rapid coating of particles by dye and reduction of removal percentage.

Determination of the optimal conditions
Finally, the optimum point for the independent variables can be calculated using Eq. ( 27) which is presented in Table 6.
The last row of the table is the experimental adsorption value performed at predicted optimum values of independent variables.It is evident that predicted value is very close to the experimental one at optimum point.Table 7 presents comparison of the adsorption efficiency of MB using HCMM composite obtained in this work with the data gathered from the literature.As it can be seen, the synthesized composite reveals comparable performance with other adsorbents as well as having low cost of synthesis

Results of the study of adsorption isotherm models
In order to study the adsorption isotherms related to the synthesized adsorbent and to determine the maximum adsorption capacity of methylene blue, equilibrium data were assessed using Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich (D-R) adsorption isotherm models.
The parameters and correlation coefficients of each model are calculated and reported in Table 8.All parameters except initial concentration of dye were set at their optimum values obtained in optimization section.Solutions with different concentrations of dye were prepared and removal of dye by HCMM adsorbent was performed until no changes in dye concentration within aqueous solution were observed.Then, adsorbent material was taken from the solution and was dried in an oven at 90 °C for 24 h.After that, the weight of adsorbent + adsorbed material was measured and the amount of adsorbed material on the adsorbent was calculated according to the initial weight of the adsorbent.Equilibrium uptakes were calculated by dividing adsorbed amount  onto adsorbent mass.Also, equilibrium concentration of dye in aqueous solution was measured by spectrophotometry according to the method described in Material and Methods section.
Experimental results were applied for finding parameters of isotherm models.Figure 6 (a-d) shows curve fitting of isotherm experimental data at 32 °C and pH of 8 based on Eqs. ( 8)-( 11) which are linear forms of Langmuir, Freundlich, Temkin, and D-R isotherm models in liquid phase, respectively.According to the results, it is evident that Langmuir model with a correlation coefficient of 0.984 fits experimental data more accurately than other isotherm models.Accordingly, it can be concluded that the adsorption of MB happens within homogeneous monolayer surfaces and after adsorption, no interaction exists between adsorbed dye and adjacent adsorbent surface sites (Senthil Kumar et al. 2014;Sahu & Singh 2019).According to Freundlich model, value of n above 1, reveals that adsorption is favorable (Pan et al. 2017).Obtained value for energy of adsorption using D-R isotherm method is 11.6 kJ/mol which indicates chemisorption process (Amin et al. 2015;Pan et al. 2017;Togue Kamga 2018).

Results of thermodynamic study
Figure 7a shows profile of enthalpy of adsorption of MB by HCMM adsorbent at 32 °C which is calculated based on Temkin and D-R isotherm models.It is evident that results with both models are in good agreement with each other and also the adsorption of MB on HCMM is endothermic.Also, Fig. 7b reveals profile of Gibbs free energy of adsorption which is always negative at 32 °C.Therefore, it can be concluded that the adsorption of MB on HCMM is spontaneous and endothermic.This is a general conclusion for the investigated temperature range in this study, because density of porous HCMM was estimated about 630 kg/m 3 in BET results section assuming that MM nanoparticles do not have significant effect on the density of adsorbent and therefore heat of fusion of MB and density of composite adsorbent are high enough to make enthalpy of adsorption always positive and Gibbs free energy of adsorption always negative.

Results of kinetic study
According to the results given in Fig. 8a and Table 9, curve fitting by LDF model is not satisfactory while coefficient of determination is low and uptake changes do not have linear relationship with uptake amounts.Also, regarding correlation coefficient of the pseudo-second-order kinetic model given in Table 9 and according to the exact fitting of experimental results shown in Fig. 8b, it can be concluded that this model exactly fits experimental data and offers more satisfactory results.It is well-known that following the pseudosecond-order kinetics is the characteristic of chemisorption process (Agbovi & Wilson 2021) and confirms results of isotherm study.For the intraparticle diffusion model, regarding Fig. 8c, multi-step adsorption is evident.The first step shows adsorption of dye in macropores which is rapid step, while the second step is due to diffusion in micropores (Pan et al. 2017).Also, based on the results of Table 9, intercept of second step of adsorption is higher which indicates higher thickness of microporous adsorption boundary layer and higher resistance against adsorption (Mane et al. 2007).

Conclusions
Removal of environmental pollutants such as dyes from wastewater is a challenging issue.As literature survey illustrated, adsorption is a promising method for this purpose and utilization of adsorbents based on agricultural wastes has economic and environmental benefits.In this study, holocellulose nanofibers were extracted from milled and powdered barley straw with different mesh sizes.Different analysis methods showed that when a smaller mesh size was used for cellulose extraction, high quality holocellulose was synthesized in terms of its higher effective specific surface area.After extraction of holocellulose, a nanocomposite was prepared by uniform dispersion of montmorillonite clay nanoparticles on prepared holocellulose.According to the SEM analysis results, the preparation of Holocellulose-Montmorillonite composite was successful because nanosized clay particles were distributed evenly within holocellulose nanofibers and increased specific surface area of holocellulose.The prepared composite was applied for the removal of methylene blue dye from synthetic aqueous solutions.To find optimal values of the effective operating parameters including solution temperature and pH, initial dye concentration, and mesh size of barley straw powders for the highest removal of methylene blue, response surface methodology was employed.In order to decrease Linear driving force k 1 (min −1 ) 0.0604 q e,calc (mg g −1 ) 22.39 q e,exp (mg g −1 ) 54.28 R 2 0.927 Pseudo-second-order q e,calc (mg g −1 ) 55.57 k 2 (g.mg −1 min −1 ) 0.018 R 2 0.999 Intra-particle diffusion k 3,1 (mg g −1 min −0.5 ) 1.766 I 1 (mg g −1 ) 39.36 R 2 1 0.992 k 3,2 (mg g −1 min −0.5 ) 0.131 I 2 (mg g −1 ) 52.96 the number of experiments, experimental design was performed using central composite design method.Based on the experimental results, an empirical second-order model was achieved which was then used to find the optimum values of each parameter.The optimum point was tested again to validate theoretical result of the model.The results revealed that HCMM composite eliminated more than 95% of methylene blue from the solution at optimum conditions.It was shown that by increasing temperature, the adsorption increased and the optimum point was achieved around 32 °C.Also, investigation of the effect of pH showed that the highest adsorptions values were achieved at near neutral to alkaline pH values, i.e., 8. Modeling of the MB adsorption process on HCMM with different isotherm models revealed that it follows Langmuir isotherm model which means that the absorption process is monolayer.Results of thermodynamic study revealed that the process is spontaneous and endothermic and while large values of enthalpy of adsorption were obtained, chemisorption takes place.Finally, kinetic study showed that pseudo-second-order kinetic model best fitted experimental results which also confirms that chemisorption takes place at the surface of the adsorbent and validates results of isotherm and thermodynamic studies.

Fig. 4
Fig. 4 Comparison of experimental values of MB removal efficiency (%) with predicted values by RSM model

Fig. 5
Fig. 5 Contour plot for binary interaction effects on adsorption efficiency of MB using HCMM composite, a temperature vs pH, b temperature vs. initial dye concentration, c temperature vs mesh size, and d initial dye concentration vs mesh size

Fig. 6
Fig. 6 Linear plots of a Langmuir, b Freundlich, c Temkin, and d Dubinin-Radushkevich isotherm models, at 32 °C and pH of 8

Fig. 7 a
Fig. 7 a Enthalpy of adsorption of MB on HCMM adsorbent at 32 °C and pH of 8 calculated based on Temkin and D-R isotherm models, b Gibbs free energy of adsorption of MB on HCMM adsorbent at 32 °C and pH of 8

Fig. 8
Fig. 8 Linear plots of a LDF, b pseudo-second-order, and c intraparticle kinetic models

Table 1
Chemicals used in different experiments of the study

Table 2
Ranges and levels of input variables for experimental design based on CCD method

Table 4
Table 5 with 95% confidence level.According to the results, high F value of 218.33, low p value of about zero, and coefficient of determination larger than 99%, designate that the obtained model is statistically acceptable.Regarding significance of different factors, it can be seen from the following table that MB removal% is linearly and quadratically dependent on temperature, and mesh size, and linearly on pH, but does not linearly change

Table 5
Analysis of variance of the obtained model for the response of the amount of adsorbed MB using synthesized bio-nanocomposite adsorbent

Table 6
Optimum conditions for achieving maximum adsorption efficiency calculated by the aid of RSM model

Table 7
Comparison of adsorption efficiency of MB in this work with other works found in the literature

Table 8
Predicted values of the parameters of Langmuir, Freundlich, Temkin, and D-R isotherm models for the adsorption of MB dye using HCMM adsorbent