Locally sourced wheat straw from the Mansa district of Punjab, India was collected. Sodium hydroxide pellets (minimum assay 97.0%) were purchased from Thermo Fisher Scientific India Pvt. Ltd. and hydrochloric acid (extrapure) was from FINAR. Ethanol (absolute) and acetone (99.8%) were from Changshu Hongsheng Fine Chemical Co., Ltd. Rhodamine B (≥95%) was purchased from Sigma Aldrich, India. Methylene Blue (>90%) was purchased from CDH (Central Drug House). In this experiment, all solutions were made with deionized water. Without additional purification, analytical-grade chemicals were employed.

The synthesis of the wheat straw adsorbent materials was investigated using two methods, illustrated in Figure 1. The first method involved the physical cutting of wheat straw to 1-2 cm pieces. The cut staw was then leached in deionised water (10 g/L) for 24 h at room temperature. The leached wheat straw was then separated using vacuum filtration and dried in Jupiter Engineering Works Hot Air Oven for 24 h at 105°C. The dried wheat straw was milled using a Retsch GM200 Grindomix knife mill and was sieved through a 0.25 mm sieve using a Retsch AS200 Vibratory Sieve Shaker.

The second protocol utilised the cut wheat straw waste in method-I and milled it using a Retsch GM200 Grindomix knife mill, followed by sieving (0.25 mm) using the same vibratory sieve stack mentioned above. This fraction was subsequently leached and separated using vacuum filtration and dried in the same manner as in method-I. The leached wheat straw feedstocks obtained through the two pre-treatment protocols were pyrolyzed in a Jupiter Engineering Works single zone tube furnace under nitrogen (99.99%) at 400°C, 10°C/min, holding for 10 min before cooling to ambient temperature. The resultant carbonised straw was washed with ethanol and acetone to remove bio-oil/tar residues liberated under the pyrolysing environment (until the washings ran from yellow to colourless), followed by drying at 105°C overnight. The collected samples were designated as BC-I and BC-II.

Using a Perkin Elmer FTIR spectrometer, FTIR spectra of samples between 400 cm⁻¹ and 4,000 cm⁻¹ were obtained. Proximate analysis of the raw wheat straw feedstock was carried out using a LECO 701 thermogravimetric analyzer at ∼1.00 g scale where moisture, devolatilization, and ash were measured. Thermogravimetric analysis was carried out as follows: ambient to 107°C at 3°C/min under nitrogen, holding for 15 min (moisture phase) before heating from 107°C to 950°C at 5°C/min, holding for 7 min (volatile phase) before cooling to 600°C. This was followed by an ashing phase in air from 600°C to 750°C at 3°C/min, before cooling to ambient temperature. Fixed carbon was calculated by subtracting the final ash mass from the sample mass before combustion. Ultimate analysis of all feedstocks was acquired using a LECO Truspec CHN Combustion analyzer using sample sizes of 50.00–70.00 mg. Scanning Electron Microscopy (SEM) images were collected via a Zeiss EVO 60 instrument at 10⁻² Pa and an electron acceleration voltage of 20 kV. Powders were adhered to a coated conductive carbon tape and attached to the specimen holder. Available surface area and porosity was measured using a Micrometrics TriStar porosimeter, prior to analysis samples were degassed for 3 h at 110°C under a nitrogen capillary feed.

On a Thermo scientific UV-Vis spectrophotometer, the concentration of RhB in a solution was measured using the characteristic peak at 554 nm using a 3.5 mL quartz cuvette with a path length of 1 cm, over a course of seven concentrations. Supplementary Figure S1 shows the calibration plot for RhB concentration in a solution via UV detection, presenting a high level of experimental accuracy (R² = 0.99797). For the case of MB, the concentration of the solution was measured using the characteristic MB peak at 664 nm, over a series of dilution factors resulting in a calibration plot where R² = 0.99874, Supplementary Figure S1.

Experiments were ran at ambient temperature, where a 100 ppm stock solution of RhB and MB were diluted to different concentrations (5, 7.5, 10, 12.5, 15, and 17.5 ppm) in borosilicate vials (30 mL). The pH of each solution was monitored and adjusted for the following, 2, 4, 6, 8, 10, and 12 using 0.1 M NaOH and 0.1 M HCl using a Metrohm Eco Titrator. To evaluate the dye adsorption by the produced materials (BC-I and BC-II), the colour change of RhB and MB dye solutions were monitored while they were constantly agitated using magnetic stirring at 500 rpm. Aliqots were removed periodically for kinetic studies, the BC-I or BC-II materials were separated via microcentrifugation using an Eppendorf Centrifuge 5,425 at 9,000 rpm for 5 min. After analysis by UV-Vis, the liquid was added back into the reaction along with the separated adsorbent.

To determine the effect of adsorbent dosage, 15 mL of each dye solution was treated with 2, 4, 6, 8, 10, 12, and 14 mg of RhB adsorbent, whereas 2, 3, 5, 6, 7, 9, and10 mg for MB the same masses were used, stopping at 10 mg as no further benefit was found by using more adsorbent after 6 mg. The initial concentration of each dye was set at 10 ppm where the solution was fixed to the ideal pH = 4 for RhB adsorption, 8 mg of BC-I and BC-II adsorbent were added, and the reaction time range was set from 5 to 1,350 min for BC-I and 5–940 min for BC-II. In contrast, MB was adjusted to optimum pH = 8, 6 mg of adsorbent (BC-I and BC-II) was added, the initial concentration was 10 ppm and the reaction time range was set from 5 to 490 min for BC-I and 5–410 min for BC-II, respectively. Eq. 1 expresses the dye removal rate: D y e   r e m o v a l   r a t e = C o   –   C e C o × 100   % (1) where C_o denotes the initial dye concentration of the solution and C_e denotes the dye concentration of the solution at adsorption equilibrium.

The surface area, pore volumes, and pore diameters of BC-I and BC-II were obtained using the BET (Brunauer-Emmett-Teller) method and the BJH (Barrett-Joyner-Halwnda) method, respectively, as shown in Supplementary Figure S2; Table 2. The isotherm stacked plots shown in Supplementary Figure S2 can be categorized as Type IV. Mesoporous adsorbents exhibit type IV isotherms and feature pores with a size range of 2–50 nm. The surface areas of BC-I and BC-II determined using the BET method were 6.284 m²/g and 20.754 m²/g, respectively. The inset image in Supplementary Figure S2 shows the BJH plots for both BC-I and BC-II, these clearly indicate that pores present are largely below 5 nm, where the average pore sizes shown in Table 2 are 9.96 nm and 7.78 nm for BC-I and BC-II, respectively. This means that the pre-treatment protocol has developed two materials of differing available surface areas and porosities, as BC-II was found to have a pore volume 3.5x higher than BC-I.

Table 3 presents the physical properties of the feedstocks and adsorbent materials. The choice of an effective thermal conversion method is significantly influenced by the moisture content. For a process like a pyrolysis to function properly, the biomass moisture content must not be more than 10%–14% (Danish et al, 2015). The fact that the unleached and leached samples have a moisture content ranging between 5% and 7% demonstrated the suitability of the raw materials for the manufacture of biochar. The ash content of the wheat straw was found to be naturally higher than other reports (Arvelakis et al, 2001; Skoglund et al, 2013; Huang et al, 2017). This value was substantially driven down by water washing (leaching), a technique known for the extracting and solubilising water soluble elements such as K, Na, Cl, Mg and to a lesser degree, Ca, S, and Si which often require elevated temperatures for effective removal (Taylor et al, 2019; Alabdrabalameer et al, 2020; Taylor et al, 2020). Table 3 shows that there is a slightly higher ash value for the second pre-treatment protocol. However, this value could infact be similar as the first pre-treatment protocol due to a large amount of fixed carbon present in the leached (method-I) sample. The purpose of removing the ash from the wheat straw was to develop the surface area of the material for the pyrolysis step by removing non-carbon entities which could promote or hinder pyrolysis reactions. An example is that alkali metals have been found to previously promote char degradation reactions at high temperatures (Saddawi et al, 2012). Additionally, metal sites could inhibit the adsorption of dye molecules or in contrast selectively chelate with the organic molecules (Rana et al., 2018). Unless impregnated directly with a known concentration of such a metal, ash content would act as a rogue entity where adsorption could not be effectively controlled. Interestingly, the leached (method-I) sample that possessed a high fixed carbon content (Table 3) was also the sample which, after pyrolysis exhibited a diminished surface area and pore volume (Table 2). Table 3 also shows the CHN values of all the materials, after the leaching process there has been an increase in carbon due to a reduction in ash. Interestingly, the carbon value for both leached materials is similar to one and other, where the method-I material has a reduced nitrogen content. Only BC-I shows a mild increase in carbon content from its pre-pyrolyzed counterpart 44.9% and 51.4%, respectively. Whereas the BC-II material has a very similar carbon content to its parent feedstock, 43.3% and 44.1%, respectively. With this in mind, due to the low energy thermal processing method used (400°C), the BC-I and BC-II materials, although lower in ash still contain a high oxygen content 43.80% and 53.00%, respectively. Typically, higher pyrolysis temperatures are used to remove oxygen from lignocellulosic matter.

The weight loss and derivative weight loss curves (Supplementary Figures S3, S4 respectively) show stacked plots for both the leached materials. Here one can observe the subtle differences between the wheat straw post leaching, using the two protocols. Before carbonization, samples were dried, as a result the samples presented a small weight loss during the moisture phase, illustrated in Table 3. As the temperature increased, there was a sharp weight loss between 215 to 360°C, this is the main volatile region. The greatest mass loss happened at 215–360°C, with a loss of 53.55% and 53.18% for method-I and method-II, respectively as shown in Supplementary Figures S3, S4.

Figure 2 depicts the SEM images of the wheat straw feedstock before leaching, after leaching and after pyrolysis. Electron microscope reveals the surface structure of biochar and gives an elaborate elucidation regarding the dispersion of mesopores in the post pyrolysis materials (Yaashikaa et al, 2020b). The surface topography between the raw and leached wheat straw (Figures 2A, B vs Figures 2C, D, respectively) show that the pre-treatment has caused some mild disruption to the ordered structure by causing some breakages and separation to the wheat straw. Meanwhile, following thermal processing at 400°C, the temperature at which indicated by thermogravimetric analysis that the major volatile components should now be removed. The thermal processing has resulted in a very distorted surface morphology for both BC-I and BC-II, Figures 2E–H, respectively. Figure 2E shows how the ordered structure presented in Figure 2B has been fractured, folded in on itself (Figure 2F) and segments are seen to be separating from the parent grain. Figure 2G also shows the end of a wheat straw grain and its natural pore system (Figure 2H). The temperature used was not high enough to cause dramatic alterations to the ends of the straw sections, instead the low pyrolysis temperature appears to have only impacted the outer surface of the straw, increasing the surface area and porosity of both the carbonised materials, mildly (Supplementary Figure S2; Table 2).

Fourier Transform Infrared Spectroscopy spectra of the raw, leached (method-I), leached (method-II), BC-I and BC-II biomass waste samples are shown in Figure 3. The spectra of the samples disclosed several functional groups that may be in charge of the dye adsorption process. The hydroxy (-OH) stretching vibrations, characteristic of moisture is identified at 3,400 cm⁻¹ (Chen and Li, 2020). The peak at 1,616 cm⁻¹ are known to be C=C and C=O stretches (Zhao et al, 2021). Both raw and leached materials showed significant absorbance peaks at 1,382 cm⁻¹, which were are identified as C-H bending vibrations (Mupa et al, 2016). At 2033 cm⁻¹, there was additional evidence of C-H stretch. Primary alcohol stretching vibrations, present in the lignin sections of the sample are responsible for the peak at 1,100 cm⁻¹ (C-O). The peak at 613 cm⁻¹ is characteristic of the silicate minerals (Si-O-Si) found in biochar samples, specifically from the retained ash. Peaks before and after leaching were nearly the same. Method-II material showed an extra peak at about 1730 cm⁻¹, which is identified as lignin and C=O unconjugated hemicellulose stretching vibrations, the latter present due to the low pyrolysis temperature used (Moharm et al, 2022). For BC-I, the O-H group from adsorbed water is still situated at 3,415 cm⁻¹, this region appears more intense than previous samples due to the natural hygroscopic nature of biochars (Chen and Li, 2020). The vibrations of C-O stretching in cellulose and hemicellulose are visible in the band at 1,102 cm⁻¹, again present due to the low pyrolysis temperature used which infers that full devolatilization has not taken place. The stretching vibration of C=C of aromatic components and lignin in the BC-I and BC-II is what created the peak at 1,624 cm⁻¹ (Zhu et al, 2014; Li et al, 2016; Janu et al, 2021). The peak of BC-II at 1733 cm⁻¹ is also identified as C=O (COOH) vibrations in lignin (El-Sayed, 2011).

A series of dye adsorption tests were performed to assess the viability of carbonized wheat straw materials for wastewater treatment, specifically investigating the effect of solid pre-treatment protocols on the produced material after thermal processing. To mimic real world water ways and treatment facilities, the adsorbents were continuously stirred into the aqueous RhB and MB solution at room temperature, opposed to a static system. By agitating the suspension, diffusion of the dye molecule into the bio renewable adsorbate as well as diffusion of the material throughout the water sample can be achieved, preventing mass transfer limitations. The mixtures were sampled periodically where after centrifugation (9,000 rpm), the supernatant was utilized to detect the absorbance at a maximum wavelength (λmax = 554 nm and 664 nm) to determine the concentrations of RhB and MB dyes, respectively. After analysis, the adsorbate and adsorbent were returned to the stirred mixture.

The extraction of RhB and MB were investigated using two alternative kinetic models: pseudo-first-order and pseudo-second-order models. They were assessed for rate constants, adsorption quantity, and correlation coefficient (Vahidhabanu et al, 2019). The premise of the pseudo-first-order model is that the rate of change in solute adsorption is proportional to the difference between saturation concentration, and the volume of adsorptive material adsorbed over time. According to the Pseudo-second-order kinetic theory, interactions such as ion sharing and transferring among the adsorbent and adsorbate control the adsorption rate. In general, the Pseudo-first-order model demonstrates the presence of physical adsorption and the Pseudo-second-order model demonstrates chemical adsorption.

Pseudo-first-order and pseudo-second-order models were investigated using Eqs 2, 3, respectively, to comprehend the adsorption phenomenon; log   Q e − Q t = log ⁡ Q e − K 1 t 2.303 (2) t Q t = 1 K 2 Q e 2 + t Q e (3) where, Q_e and Q_t represent the amount of dye adsorbed (mg/g) at equilibrium and at a time, t (min.), respectively, K₁ and K₂ represent the rate constant for pseudo-first-order (min⁻¹) and pseudo-second-order (g mg⁻¹ min⁻¹) adsorption, respectively (Xiao et al, 2018; Ganguly et al, 2020).

The parameters found by fitting the experimental data to the kinetic Eqs 2, 3 are reported in Table 4 along with the fitting outcomes in Supplementary Figure S5. The correlation coefficient value must be around 1, and the adsorption quantity (Qe) value must be close to the experimental value, depending on whether the adsorption is well suited for pseudo-first-order or pseudo-second-order (Yaashikaa et al, 2020b). In contrast to the pseudo-first-order equation, which offered excellent fits to the experimental data, the pseudo-second-order model for BC-I and BC-II did not perform well. For the Pseudo-first-order model, the theoretical adsorption quantity is in good accordance with the experimental adsorption quantity for both the dyes by BC-I and BC-II. For BC-I, the correlation coefficient values for the Pseudo first-order model are 0.9545 for RhB and 0.9895 for MB. For BC-II, the correlation coefficient values for the pseudo-first-order model are 0.9989 for RhB and 0.9933 for MB. As a consequence, the system adheres to a pseudo-first-order kinetics model with high correlation coefficient values and adsorption quantity for both RhB and MB by BC-I and BC-II. This is due to the physical adsorption that takes place between dyes and the carbonized wheat straw adsorbents in water.

An isotherm is a curve that correlates an adsorbate’s equilibrium concentration with its capacity to adsorb at room temperature, and it can be used to illustrate how adsorbate and adsorbent interact.

Two isotherm models, including the Langmuir and Freundlich models were experimentally analyzed. The Langmuir adsorption isotherm describes homogenous adsorption which leads to a monolayer covering the adsorbent surface with a limited number of identical sites. The Langmuir equation is represented as follows: C e Q e = 1 K L Q M + C e Q M (4)

Where C_e, Q_e, Q_M, and K_L stand for, in that order, the adsorbate’s equilibrium concentrations (mg/L), the ratio of dye adsorbed to adsorbent per gram at equilibrium (mg/g), the highest possible monolayer coverage adsorption capacity (mg/g), and the Langmuir constant (L/mg) (Fan et al, 2016).

The Freundlich adsorption isotherm shows a multilayer coverage because the adsorbates are heterogeneously adsorbing at the surface of the adsorbents. The Freundlich equation is represented as follows: log ⁡ Q e = log ⁡ K F + 1 n log ⁡ C e (5)

Where Ce, Qe, and K_F stand for, respectively, the adsorbate’s equilibrium concentration (mg/L), the ratio of dye adsorbed to adsorbent per gram at equilibrium (mg/g), the adsorption intensity, and the Freundlich constant (mg^1−1/n L^1/n g⁻¹) (El-Hosiny et al, 2018; Maruthapandi et al, 2018).

In addition to the model fits of the isotherm Eqs 4, 5 for RhB and MB by BC-I and BC-II, the parameters derived by employing these equations are shown in Table 5. The R² value of the Langmuir and Freundlich model is close to 1 as shown in Table 5. The experimental data matches the Freundlich model more closely than the Langmuir model, as shown by the correlation coefficients (R², Table 5) along with the fitting model as shown in Supplementary Figure S6. These findings disclosed that RhB and MB were adsorbed onto the BC-I and BC-II in a heterogeneous manner, ensuring the multilayer coating of dye molecules on the biochar outer surface. Adsorption is facilitated as values of “n” are greater than unity. The adsorption capacity would be impacted by the carboxyl group because it belongs to the class of functional groups in biochar (Figure 4). The cationic dyes RhB and MB may be adsorbed due to the involvement of the carboxyl group, which has a net negative charge and is possibly the major functional group.