Effective toluene removal from aqueous solutions using fast pyrolysis-derived activated carbon from agricultural and forest residues: Isotherms and kinetics study

Abstract

In this study, the production and characterization of activated carbons (ACs) from agricultural and forest residue using physical activation are discussed. Biomass-based biochars produced during fast pyrolysis process is introduced as alternative precursors to produce AC and the integrated process for the co-production of porous adsorbent materials from biochar via the fast pyrolysis process is suggested. Moderate surface areas and good adsorption capacities were obtained from switchgrass (SWG) and pine tops (PT) based AC. The surface areas were 959 and 714 m²/g for SWG- and PT-based AC, respectively. The adsorption capacities using toluene as pollutant for two model systems of 180 and 300 ppm were measured and ranged between 441-711 and 432–716 mg/g for SWG-based and PT-based AC, respectively. The nitrogen adsorptive behavior, Lagergren pseudo-second-order kinetic (PSOK) model and kinetics isotherms studies describe a heterogeneous porous system, including a mesoporous fraction with the existence of a multilayer adsorption performance. The presence of micropores and mesopores in SWG- and PT-based AC suggests potential commercial applications for using pyrolytic biochars for AC production.

Graphical abstract

Highlights

• •Fast and efficient toluene removal was achieved using pyrolytic biochar.
• •Ag. and forest residues AC shown strong adsorbate-adsorbent interactions with toluene.
• •Possible adsorption mechanisms were supported by kinetic and isotherms studies.
• •Adsorption capacities of activated pyrolytic biochar was compared with commercial AC.
• •Pyrolytic biochar offers a real opportunity for biomass-based adsorbents towards aromatics removal.

Nomenclature

Abbreviation – Definition

AC

Activates carbon

ACUS

The activation unit system

ASTM

American Society for Testing and Materials

AW

Acid washing

AWAC

Acid washed active carbon

BET

Brunauer Emmett Teller

C

Carbon

CHN

Carbon, hydrogen, and nitrogen

IUPAC

International Union of Pure and Applied Chemistry

LCA

Life cycle assessment

PSD

Pore size distribution

PSOK

Lagergren Pseudo-Second Order Kinetic

PT

Pine tops

SWG

Switchgrass

TAPPI

Technical Association of the Pulp and Paper Industry

TEA

Techno Economic Analysis

USD

United States Dollars

U.S EPA

United States Environmental Protection Agency

UV-VIS

Ultraviolet–visible spectroscopy analysis

1. Introduction

The global market for activated carbons (ACs) has grown steadily and was valued at USD 5.7 billion in 2021. Demands for AC are projected to reach USD 8.9 billion by 2026 according to a recently published market report for AC [1,2]. One of the major drivers for this growth is the usage of AC for the removal of organic and inorganic pollutants in wastewater treatment. To date, AC has been mainly utilized in industrial and municipal wastewater plants, which accounts for 45% of the AC usage [3]. However, in North America, demands for AC are projected to constantly increase due to population growth, additional needs for oil and gas production, and the pharmaceutical industry [1].

AC is defined as a carbonaceous material [4], with high surface area [5,6], well-developed structural porosity [[7], [8], [9]], high adsorptive capacity [10] and physicochemical stability [11]. Resources to produce AC can vary widely and includes coal, bio-waste material, agricultural residues, and forestry biomass. To produce AC from different resources, each resource requires different activation conditions, and the resulting AC has characteristic performances and different cost parameters [10,[12], [13], [14]].

Thermal treatment (physical activation) consists of two stages, where both stages take place separately. In the case of biomass, carbonization occurs first to convert the oxygen rich biomass into a carbon rich biochar. Typically, this carbonization is conducted under an inert atmosphere using N₂ or CO₂ rich gas at temperatures between 400 and 600 °C, followed by activation of the biochar at temperatures between 600 and 850 °C [4]. The aim of the carbonization is to remove the oxygen that makes up 40–45% of the original biomass and generate some initial porosity in the resulting carbonized material. During carbonization, the biomass loses 40–60% of its mass originating from the loss of oxygen compounds and the remaining carbon atoms are arranged in a loose graphitic-like structural. Therefore, the resulting biochar has a relatively low surface area [15]. Prior research has shown that carbonization is dominated by the loss of oxygenated compounds by converting them into the gaseous form via the decomposition of hemicelluloses and pyrolysis of cellulose and lignin fractions [[16], [17], [18], [19]]. Afterward, subsequent reactions in the presence of activating agents, such as steam, CO₂, air or some combination of these oxidizing agents [20] led to the release of carbon oxides from the carbon surface [21], causing in some cases additional condensation reactions [4]. The activation step induces reactions of the remaining oxygen components present in the carbonized material. During the activation, both carbon monoxide and carbon dioxide are generated, resulting in the creation of new pores and increase in the surface area [15,22]. The development of physical pores during the activation process can be classified into three phases: creating channels that connect previously inaccessible pores, developing new pores, and widening of the existing pores [23]. The characteristics of pores created during the activation can be varied by the types of biomasses, reaction temperature, time, and inert gas employed for the activation. However, mechanisms still lie in a black box and further studies are required.

Chemical activation, also known as wet oxidation, differs from physical activation. This is because this process can conduct the carbonization and activation simultaneously. During the chemical activation, biomass is subjected to thermal treatment and activation at the same time [[24], [25], [26], [27]]. As the name implies, chemical activation requires a reactant that is impregnated with the raw biomass (by soaking) to help generate the AC [[27], [28], [29]]. Chemical activation generally required lower reaction temperatures compared to physical activation. Commonly, the temperature ranges for chemical activation are between 300 and 700 °C, versus 600–800 °C for physical activation [[30], [31], [32]]. The properties of AC produced by chemical activation are determined by the reactivity of employed inorganic reactant and their ability to create oxygenated functional groups into the surface of AC during the activation process. Typically, reactants such as H₃PO₄ [33], ZnCl₂ [[34], [35], [36]], H₂SO₄, KCNS [37], H₂O₂, KMnO₄ [24], KOH [38], and NaOH [39] can be employed for the chemical activation.

Aromatic compounds (e.g., toluene, benzene) are widely employed in various industries as organic solvents and raw materials. However, aromatic compounds have a high toxicity, flammability, and carcinogenic nature [40,41]. Furthermore, it is hard to remove aromatic chemicals once they are released into the ecosystem. Therefore, proper treatments should be done to industrial effluent before discharging to ensure the safety of the overall ecosystem and water supply. Extensive studies have been conducted to verify and validate the utilization of AC as an adsorbent to treat aromatic compounds in wastewater. The validity of adsorption using AC has been proven and renowned as one of the best available environmental control technologies for pollutants removal on gaseous and liquid phases [[42], [43], [44]]. As for organic compounds in the liquid or gas phase, the adsorption capacity of AC is closely related to two factors: the volume of narrow micropores in AC and the presence of functional groups that can contribute to the adsorption of organic compounds. Some authors have also noted that the mesoporous fraction (2–50 nm, IUPAC) of the AC contributes to the adsorption of these aromatic species [[45], [46], [47], [48]]. Both possibilities are feasible considering the different experimental conditions employed for these studies. In addition to the adsorption capacity, the adsorption rate is another important factor that impacts the feasibility of AC usage as an adsorbent for wastewater treatment [49,50].

It is worth mentioning that the availability of biomass used for AC production is typically determined by the price and potential activation extent [51]. There have been strong desires to seek cost-effective alternative sources and processes to produce AC [8]. A huge amount of agricultural and forest residues is produced as by-products in intermediate processing facilities and these residues have drawn attention due to their high carbon contents, low inorganic contents, and cost effectiveness [52]. So far, utilization of these by-products has been limited. Some of them are simply burnt to generate heat and steam, but most of them have been just disposed of, resulting in incurring additional costs. However, the agricultural and forest residues meet basic the conditions to be AC and these advantages make agricultural and forest residues emerged as attractive resources for AC production [[53], [54], [55]]. For these reasons, the study of pyrolytic biochars produced from agricultural and forest residues as an alternative precursor for the production of AC and their possibility as an adsorbent for the removal of aromatic compounds are presented in this work.

In this work, co-located activated carbon (AC) production is suggested to be a promising procedure for producing high-performance adsorbent materials. This co-location process is proposed as an extension of the fast pyrolysis process, using the resulting pyrolytic biochars as a precursor for AC production. In this study, one agricultural residue (switchgrass, SWG) and one forest residue (pine tops, PT) were selected to produce AC owing to their vast availability, cost-effectiveness, and year-round supply. To characterize ACs produced from different biomass resources, properties of ACs closely related to adsorption were investigated: surface area, pore size distribution, pore volume and width, particle size, chemical textural morphology, and solid surface acid-base characterization [[56], [57], [58]]. Afterward, the adsorption performances of these ACs were tested using toluene as a model pollutant in wastewater. In commercial practice, toluene is often included in industrial wastewater ranging from detectable levels from 180,000 to 250,000 μg/L (180–250 ppm), and at much lower levels, e.g., 237 μg/L (0.2–0.3 ppm) in municipal sewage [59,60]. Therefore, adsorption experiments were designed accordingly. After the adsorption, several different isotherm and kinetic models were employed to interpret the adsorption behaviors of ACs. This study reveals the great potential of using agricultural and forest residue such as SWG and PT as precursors to produce AC and helps expand the usage of the fast pyrolysis process to convert rejected biomass into commercially valuable AC for the removal of aromatic compounds in wastewater streams.

2. Materials and methods

2.1. Materials

Forest and agricultural residues (pine tops and switchgrass, respectively) from southern U.S were provided by Genera Energy Inc (Vonore, Tennessee) and toluene in liquid form as received (Certified ACS) from Fisher Scientific (Hampton, New Hampshire) Assay 99.5%.

2.2. Biomass preparation

Oven-dried biomass (105 °C for 24 h) was treated by a Wiley Mill equipped with a 0.5 mm screen to help optimize the heat and mass transfer during both carbonization and activation processes. Afterward, the biomass in powder-like form was stored in plastic bags to prevent the adsorption of moisture. The moisture contents of biomass ranged from 5 to 10%

2.3. Biomass chemical analysis

2.3.1. Composition analysis

The contents of extractives were measured employing the TAPPI Standard T204cm-07 [61]. After air-drying biomass, 2 g of air-dried biomass was placed in an extraction thimble. A solvent for the extraction was prepared with benzene and ethanol (v/v = 2:1) and the extraction of extractives in biomass was performed with Soxhlet extraction apparatus for 24 h. The content of extractives was computed using Equation (1) as follows:

$$Extractives\left(wt.\%\right)=\left[\frac{W_e-W_h}{W_p}\right]\times 100$$ (1)

Where, W_e is vacuum-dried weight of extractives (g), W_h is vacuum-dried weight of biomass (g) and, W_p is vacuum-dried weight of blank residue (g).

The composition of biomass and ash content were investigated with extractive-free biomass by following a procedure described elsewhere [62]. Briefly, 300 mg of oven-dried (105 °C, overnight) biomass were treated with 3 mL of 72 wt % sulfuric acid for 1 h, followed by dilution to 3 wt % sulfuric acid. Then, the hydrolysate was autoclaved at 121 °C for 1 h. The acid-insoluble lignin (Klason lignin) fraction was collected on a pre-weighed Gooch crucible filter (medium pore size). The content of acid-insoluble lignin was gravimetrically measured by comparing the weight difference before and after the oven-drying by following Equation (2). After the filtration, the acid-soluble lignin fraction was determined by UV absorbance measured at 205 nm using a UV-VIS spectrophotometer (GENESYS Spectrophotometer, Thermo Fisher Scientific, Waltham, MA), considering the dilution [63].

$$Acidsolublelignin\left(wt.\%\right)=\left[\frac{D\times V\times A}{a\times W}\right]\times 100$$ (2)

Where, D is dilution ratio, V is the volume of an aliquot of hydrolysate (L), A is the UV absorbance measured at 205 nm, a is absorptivity of the lignin (110 L g⁻¹cm⁻¹ at 205 nm), and W is the oven-dried weight of the biomass (g).

To measure the carbohydrate composition of biomass, the filtered hydrolysate was neutralized with calcium carbonate and the mixture was filtered through a 0.2 μm syringe filter. Then, the sample was injected to HPLC (Agilent 1200 series, Agilent, Santa Clara, CA) equipped with a Shodex SP-0810 column. The ash content of biomass was gravimetrically determined the weight difference before and after the combustion at 575 °C for 6 h.

2.3.2. Elemental and proximate analysis

Elemental analysis of biomass was performed according to ASTM-D5291 using a PerkinElmer CHN Elemental Analyzer (2400 Series II). Approximately, 6 mg of biomass was subjected to the analysis and the contents of carbon, nitrogen, and hydrogen were determined by a pre-developed quantitative method installed in the instrument. Only the content of oxygen was computed by the weight difference before and after the measurement, trace elements (e.g., sulfur, potassium, phosphorus, and chlorine) are not quantified in this analysis. The proximate analyses were performed according to ASTM-D1762-84 using a furnace and monitoring the temperature through a thermogravimetric analyzer (Q500, TA instruments, New Castle, DE).

2.4. Activated carbon production

A lab-scale biochar reactor was used for both the carbonization and activation processes. The activation unit system (ACUS) had an isolated heating chamber to provide heat and maintain the temperature of the cylindrical stainless-steel reactor constantly. A sample holder with approximately 350 g of biomass was placed into the cylinder vessel. The carbonization and activation were processed at 500 and 800 °C, respectively. To maintain oxygen-free environment of the ACUS during the carbonization (35–40 min) and activation (1 h), nitrogen gas was injected to the instrument at the flow rates of 4 L/min and 2 L/min, respectively. The steam was generated employing an isocratic pump with a heated chamber held at 500 °C. After cooling at room temperature (23 °C), the resulting carbonaceous materials were collected. The yields of biochar/activated carbon were gravimetrically calculated.

2.4.1. AC acid washing

Acid washing (AW) of the AC was performed to dissolve inorganic foulants such as calcium carbonates, magnesium and sodium salts, fine minerals such as silica and iron particles, and some pyrolytic condensates, that potentially could obstruct the porous network. The AW was carried out by adapting previous protocols described elsewhere [[64], [65], [66]]. Briefly, activated carbon (2 g) was soaked into 40 mL of 3 wt % HCl acid aqueous solution at room temperature for 5 h under continuous stirring. Then, the resulting acid washed samples were filtered, rinsed with deionized water, and dried in an oven at 105 °C for 12 h to finally obtain the acid washed AC (AWAC).

2.5. Biochar and activated carbon characterization

2.5.1. Brunauer Emmett Teller (BET) surface area and micropore volume analysis

BET analysis was performed in order to analyze the surface area and micropore volume of the resulting biochar and activated carbons, using a Gemini VII 2390 surface area analyzer (Micromeritics, Norcross, GA). A sample (approximately 0.10–0.20 g) was loaded into a quartz-bulb tube probe, degassed for 2 h at 200 °C and placed into the BET unit. The BET surface area was estimated by multipoint calculations based on the linear relative pressure regime from 0.05 to 0.30 (p/po) [67]. Total pore volume was obtained at p/po = 0.98. Finally, microporous, mesoporous, macroporous estimation is obtained by the summation of pore volume diameters of less than 2 nm, 15 nm, 1000 nm respectively, as well as pore size distribution [68].

2.5.2. Boehm's titration

The selective neutralization method was used to determine the amount of acidic surface oxygen groups (phenolic, carboxylic, and lactonic groups) using Boehm's titration method. This method provided the overall measurement of the functional groups present on the solid surface of the AC. First, AC (0.5 g) was placed in different flasks containing 50 mL (0.05 N) of sodium bicarbonate (NaHCO₃), sodium carbonate (Na₂CO₃), sodium hydroxide (NaOH), to determine the acidic oxygen surface functional groups and hydrochloric acid solution (HCl) for basic surface groups estimation. The flasks were then sealed and stirred for 24 h at room temperature. Then, ACs in the flasks were collected via vacuum filtration and 10 mL of filtered solution was titrated with 0.05 N sodium hydroxide or hydrochloric acid, depending on the initial solution utilized for the analysis.

2.5.3. Batch-wise adsorption chamber

To investigate the adsorption isotherm and kinetics, a non-conventional technique was designed due to the high volatility of toluene. A syringe-like chamber (Fig. 1) was used to eliminate the headspace typically found in open adsorption systems. The fresh adsorbent (AC) was placed inside the 50 mL syringe with a magnetic stirrer utilized for non-flat surfaces, later 50 mL of water/toluene solution was pulled back into the syringe. The tip of the syringe was connected to a 1.5 μm disposable syringe filter and the syringe filter's outlet was sealed with a film to prevent solvent evaporation. Samples were taken at predetermined time intervals by gently pressing the plunger of the syringe. To avoid the generation of head space, samples (2 mL) were collected in a 2 mL vial. Afterward, a UV-VIS spectrophotometer (GENESYS Spectrophotometer, Thermo Fisher Scientific, Waltham, MA) was employed to measure the concentration of toluene at the wavelength of 268 nm (Schonherr et al., 2018). For all the adsorption isotherms, experimental data for two toluene concentrations (180 and 300 ppm) were evaluated in a room-controlled temperature (23 °C). Adsorption isotherms and kinetic experiments were conducted at the appropriate dosages. This protocol allowed for the assessment of different parameters such as solution concentration, adsorbent dose, contact time, adsorption temperature, etc. under controlled time and temperature. The adsorption capacity of the activated carbon at a given time (q_t) was evaluated using Equation (3) as follows:

$$q_t=\frac{\left(C_o-C_t\right)}{\raisebox{1ex}{$W$}\!\left/ \!\raisebox{-1ex}{$V$}\right.}$$ (3)

Where C₀ and C_t are the respective concentrations of toluene (gr/L = ppm) in solution at time zero and on a given time t, respectively. W is the mass of the adsorbent; V is the volume of the solution.

Fig. 1.

Syringe-like set up for batch adsorption (Adap. Asenjo).

2.5.4. Adsorption kinetics

The first and second order Lagergren model was used to represent the experimental results. The rate of change in adsorbate uptake with time was directly proportional to the difference in saturation concentration and the amount of solid uptake with time applicable over the initial stage of an adsorption process by following Eqs. (4), (5), (6), (7)) respectively and thoroughly adjusting q_e as suggested by Ranjan [69]. The Lagergren pseudo–first-order model can be described by the following linear form:

$$\frac{d_{qt}}{d_t}=k_{eq}\left(q_e-q_t\right)$$ (4)

where q_e and q_t (mg/g) are the amounts of the amount adsorbed at equilibrium and at time t, respectively; k_eq is the equilibrium rate constant in the pseudo–first-order model (L/min).

After integration and applying the boundary conditions, the following equation is obtained:

$$\log \left(q_e-q_t\right)=\log \left(q_e\right)-\frac{k_{eq}}{2.303}t$$ (5)

Value of k_equ is found from the linear plots of log (q_e-q_t) versus time. If an adsorption process follows true first-order kinetics, then the intercept of log (q_e-q_t) versus t plots would be equal to log of experimentally determined q_e or adjusted to fit the experimental value. For slow adsorption processes, it is hard to reach a true equilibrium, making it difficult to measure q_e accurately. Therefore, the Lagergren pseudo-first-order equation may not fit experimental data for the whole range of adsorption times, for extended times. It is usually applicable over the initial 20–30 min of the adsorption process and it is then necessary to extrapolate the experimental data to infinite time to obtain q_e, which is often not possible. In this case, q_e is an adjustable parameter to be calculated.

In some cases, the pseudo–first-order equation is expressed as,

$$\frac{1}{q_t}=\frac{K_{ads}{Q}_1}{t}+\frac{1}{Q_1}$$ (6)

where Q₁ is the equilibrium adsorption capacity (mg/g), qt is the adsorption capacity at time t (mg/g), and K_ads is the adsorption rate constant of pseudo–first-order adsorption (min⁻¹). If the adsorption process is in accordance with this equation, 1/q_t will show a perfect linear relationship with regard to 1/t.

The pseudo–second-order kinetic model assumes that the rate-limiting step is chemical sorption (chemisorption) and predicts the behavior over the whole range of adsorption. In this case, the adsorption rate is dependent on adsorption capacity not on the concentration of adsorbate. One major advantage of this model over Lagergren first order model is that the equilibrium adsorption capacity can be calculated from the model; therefore, there is theoretically no need to evaluate the adsorption equilibrium capacity from experiments []. The differential equation for the pseudo–second-order kinetics is given by

$$\frac{d_{qt}}{d_t}={k_{eq2order}\left(q_e-q_t\right)}^2$$ (7)

Several theoretical models can be used to fit experimental adsorption data. In this study a first and a second order adsorption models were used according to Liu [70], where t is time (min), q_t is the amount of sorbate adsorbed at time t (mg/g), K_eq2order is the pseudo–second-order rate constant model (g mg⁻¹min⁻¹), q_e is the amount of sorbate adsorbed at equilibrium (mg/g). For the Lagergren model, the first order of adsorption kinetics is described by the following Eqs. (8), (9)):

$$\frac{d_q}{d_t}=K_L\left(q_e-q_t\right)$$ (8)

After integration and setting the boundary conditions of t = 0 up to t, and q = 0 to q = q_t, resulting in:

$$q_t=q_e\left(1-e^{-K_{L}t}\right)$$ (9)

For the pseudo-second order kinetic model, the following expression is used:

$$\frac{d_q}{d_t}={K_{SE}\left(q_e-q_t\right)}^2$$ (10)

After integration for the boundary conditions of t = 0 up to t, and q = 0 to q = q_t, we obtain:

$$q_t=\frac{K_{SE}{q_e}^{2}t}{1+K_{SE}{q}_{e}t}$$ (11)

For both models, and both toluene concentrations the parameters K_L and K_SE were estimated by the partial least minimization of the error function Σ (qt^exp – qt^calc)² where qt^exp is the value at the experimental data and qt^calc are the resulting values by applying Eqs. (9), (11)).

Three widely used adsorption isotherms, (i.e Langmuir, Freundlich and Temkin) have been employed for modeling adsorbate-adsorbent interactions, and the potential of adsorbents. The most used adsorption isotherms are based on the non-linear form of both the Langmuir and Freundlich models, Eqs. (12), (13)), respectively. While the Temkin's non-linear form is shown on equation (14).

$$q_e=\frac{q_m{K}_L{C}_e}{1+K_L{C}_e}$$ (12)

$$q_e=K_F{C}_e^{1/n}$$ (13)

$$q_e=\frac{TR}{b}\ln \left({K_{T}C}_e\right)$$ (14)

Where, Ce (mg L⁻¹) is equilibrium concentration in solution, q_e (mg g⁻¹) is the adsorption capacity; K_L (L mg⁻¹) and K_F ((mg g⁻¹)/(mg L⁻¹)^1/n)are the Langmuir and Freundlich adsorption equilibrium constants, respectively; n is a heterogeneity factor mostly related to the intensity of adsorption. K_T is the equilibrium binding constant (L/mg) and b represents the variation in adsorption energy (kJ/mol). T represents temperature (K) and R is the universal gas constant (8.314 J mol⁻¹ K⁻¹).

2.6. Statistical analysis

All experiments were triplicated to present experimental results as mean ± standard deviations. The experimental data (for studies corresponding to results shown in Table 1, Table 2) was process via one-way analysis of variance (ANOVA) followed by Tukey tests to examine differences among mean values. If p-values <0.05, then, the mean values were considered statistically significant.

Table 1.

Chemical, elemental and ultimate analysis composition of raw biomass.

% wt. Dry basis
Component	SWG	PT
Chemical analysis
Extractives	4.7 (0.1)a	10.3 (0.2)b
Cellulose	38.5 (1.0)a	28.6 (0.1)b
Hemicellulose	28.4 (0.7)a	19.8 (0.0)b
Lignin (Acid Soluble Lignin + Klasson Lignin)	21.4 (0.1)a	37.5 (0.5)b
Acetyl	4.7 (0.1)a	1.5 (0.0)b
Mass losses	1.0 (0.1)a	1.2 (0.2)b
Elemental analysis
C	49.2 (0.0)a	53.5 (0.1)b
H	6.1 (0.1)a	6.4 (0.1)b
N	0.2 (0.0)a	0.5 (0.1)b
O	44.5 (0.1)a	39.6 (0.1)b
Ultimate analysis
Volatile matter	86.4 (0.3)a	81.5 (0.4)b
Fixed carbon	12.3 (0.2)a	17.4 (0.1)b
Ash	1.3 (0.03)a	1.1 (0.05)b

*Value in parenthesis represent the standard deviation calculated from three replicates and different letters within in a row indicate that there are significant differences (p < 0.05).

Table 2.

Textural parameters of carbon materials under study.

AC Precursor	Surface area (m2/g)		Pore Size (nm)
	Biochar	Activated	Average pore width	Average pore diameter
SWG	102.8 (6.1)a	958.8 (7.9)a	3.8 (0.7)a	11.1 (6.5)a
PT	22.9 (8.7)b	713.7 (9.4)b	6.4 (0.6)b	9.4 (3.6)a
AC-REF	–	1657.9 (11.7)c	5.7 (0.8)b	2.8 (0.9)b

*Value in parenthesis represent the standard deviation calculated from three replicates and different letters within in a column indicate that there are significant differences (p < 0.05).

3. Results and discussion

3.1. Biomass characterization

Three different characterization methods were employed to systematically analyze biomass feedstocks. The chemical composition of the initial biomass feedstocks is shown in Table 1. The contents of ash in biomass are an important factor to determine conditions of the initial pyrolysis process because alkaline earth minerals have a significant impact during the pyrolysis [71]. It was revealed that SWG contain more ash (1.3 wt %) compared to PT (1.1 wt %). The composition analysis showed that SWG contained more carbohydrate fractions (cellulose + hemicelluloses, (66.9 wt %) compared to PT (48.4 wt %). In contrast, PT had more lignin. These results are expected considering the origin of biomass. These carbohydrate fractions can be converted to anhydrosugars, organic acids, ketones, furans, and aldehydes during the initial carbonization reactions [72]. Whereas, during the initial carbonization, lignin fractions are converted into phenolic compounds, light oxygenates, and char [72]. Elemental analysis results are shown in Table 1. The carbon content is another critical parameter to produce biochars and AC [73]. The greater content of carbon was observed in PT (53.5 wt %) compared to SWG (49.2 wt %). In the cases of nitrogen and hydrogen contents, PT had more nitrogen and hydrogen, however, differences in nitrogen and hydrogen contents between SWG and PT were small. Based on the ultimate analysis, it was found that PT produced a greater amount of fixed carbon (17.4 wt %) compared to SWG (12.3 wt %). This was possibly due to the fact that PT (21.4 wt %) contained more lignin compared to SWG (37.5 wt %).

3.2. Activation process yields

Carbonization of the biomass was sequentially followed by the activation step. A series of screening reactions were conducted to optimize conditions for the activation and these process conditions (temperature, residence time, heating rate, nitrogen gas, and steam flow rate) were kept employed (800 °C, 1 h of activation, 10 °C/min, 2 L/min, and 120 mL/h, respectively) for all the rest of the work. The highest conversion yield from biochar to AC was achieved from the PT (15.5%), followed by SWG (12.3%). The higher carbon and lignin content of the PT is possibly responsible for the for the higher AC yield. The experimental yields for these ACs were at the low end of the range reported in the literature 11.5–18.8% [74]. But it is important to recognize that these AC have been produced under physical activation conditions, which are known for providing lower yields and surface areas but also involve a low cost process.

3.3. Acid washing

The acid washing procedure influenced the overall yield of ACs because it caused an additional mass loss of approximately 14%. The pH of solutions employed for acid washing ranged between 0.33-0.39 and 1.27–1.42 before and after washing, respectively. The acid-soluble ash fraction was removed along with the potential bio-oil condensates deposited onto the porous surface of the solid carbon, improving its morphology and textural characteristics as shown in Fig. 2. The acid washing induced changes in the elemental compositions of biochars and their changes associated with the activation and acid washing are shown in Fig. 3a and b. As expected, the activation caused a significant increase in the carbon content of the biochars produced from both feedstocks. Furthermore, additional increases in the carbon contents after acid washing were observed. The effect of acid washing was greater for PT-based biochar compared to SWG-based biochar. The greater increase in the content of carbon was observed from the PT-based biochar. The reduction in the oxygen contents was observed in both PT and SWG. This was probably due to the removal of oxygenated compounds initially deposited onto the solid surfaces [75].

Fig. 2.

SEM Images for unwashed (top) and washed (bottom) samples. Right: PT and left: SWG.

Fig. 3.

Elemental analysis for raw material, activated carbon and washed AC a) PT and b) SWG.

3.4. Brunauer Emmett Teller (BET)

In this work, the IUPAC definitions for the pore size distribution (PSD) were used, e.g., (micropores <2 nm, mesopores 2–50 nm, and macropores >50 nm) and BET analysis with nitrogen was used to determine the surface area characteristics, pore size distribution, and pore openings for porous solids. Based on the IUPAC isotherms classification, it was clear that the reference AC can be categorized into Type I isotherm approaching a limiting value when p/po < 1 (Fig. 4) [76]. These behaviors can be explained by the adsorption energy of the first monolayer adsorption. Also, these behaviors are characteristic of microporous adsorbent materials with an average pore diameter of less than 2 nm [77]. The ACs produced from SWG and PT exhibited combined behaviors of Type I and II isotherms. The adsorption began with fast adsorption at relatively low-pressure ranges where the micropores in ACs were fully occupied with nitrogen gas. Afterward, the knee points in the isotherm curves of SWG- and PT-based AC represented that the formation of a monolayer was achieved, followed by the initiation of multilayer adsorption after the plateau-like region at the medium pressures (Fig. 4) [78]. Type II isotherms are often observed from meso- and macroporous adsorbents with strong adsorbate-adsorbent interactions. Also, carbon materials having a high heterogeneity often show the characteristics of Type II isotherms. Table 2 shows the parameters obtained from the BET (N₂ adsorption) isotherms analysis. The physical activation process was able to increase the surface area of the raw SWG-based and PT-based AC up to approximately 10 and 30 times, respectively compared to their biochars. The average pore diameter of both SWG-based and PT-based AC were within the range of the micro- and mesopore based on the IUPAC definitions, which explained their nitrogen adsorption behaviors.

Fig. 4.

Nitrogen adsorption isotherms.

3.5. Boehm's titration

The number of functional groups containing oxygen is an important parameter to determine the adsorption behaviors of AC [79]. Based on adsorption mechanisms suggested elsewhere, toluene in wastewater seems to interact with AC via π-π electron donor-acceptor complex, π-π interactions, and electrostatic interactions between aromatic rings of toluene and oxygen-containing functional groups in AC [[80], [81], [82]]. Therefore, the total amount of oxygen-containing functional groups residing on the surface of the AC was determined by Boehm's titration method (Table 3) [83]. The total amount of acidic functional groups was measured as the sum of carboxylic, lactonic, and phenolic functional groups. Although SWG-based and PT-based ACs were found to have comparable total amounts of acidic functional groups, the ratios among carboxylic, lactonic, and phenolic groups residing in each AC were different. Basic functional groups on the surface of ACs were calculated based on the amount of HCl required to titrate the AC [83].

Table 3.

Surface oxygen-containing functional groups from Boehm titration.

Activated carbon	Functional group	Amount of functional group (mmol/g)	Total acidic groups (mmol/g)	Total basic groups (mmol/g)
SWG	Carboxylic	0.159 (0.051)	1.258	1.719
	Lactonic	0.167 (0.009)
	Phenolic	0.932 (0.087)
PT	Carboxylic	0.107 (0.012)	1.233	1.772
	Lactonic	0.147 (0.013)
	Phenolic	0.979 (0.073)
AC-Ref (Norit ®)	Carboxylic	0.191 (0.011)	1.324	1.576
	Lactonic	0.111 (0.016)
	Phenolic	1.022 (0.072)

*Value in parenthesis represent the standard deviation calculated from three replicates.

3.6. Batch equilibrium adsorption and kinetics

Batch equilibrium adsorption experiments were performed to examine the adsorption capacity of the PT-based and SWG-based AC using two different toluene concentrations and five adsorbent dosages for each AC. The adsorption performance of the ACs was assessed with toluene at concentrations similar to those found in industrial and municipal wastewater [59,60] and compared to the adsorption performance of commercially available AC. As shown in Fig. 5a and b, at all ranges of adsorbent dosages and toluene concentrations, the highest adsorption capacity was achieved from the reference AC, followed by PT-based AC and SWG-based AC, respectively. The best adsorption performances of the reference AC were possibly due to both its microporous nature and the amount of acidic functional groups on the surface. The surface characteristics of the reference AC such as surface area and pore size are more suitable for the adsorption of toluene (Table 2). Nevertheless, in Fig. 5a and b shows that PT-based AC adsorptive behavior is closer to the reference AC, suggesting the remarkable potential of this precursor material for toluene adsorption at industrial and municipal application levels. The reference AC has a greater surface area and smaller pore size compared to biomass-based ACs. It is important to note that commercial ACs are frequently subjected to post-treatments that functionalize the surface of the ACs [84] enhancing its adsorptive behaviors. Although biomass-based ACs showed lesser adsorption performances compared to the reference AC, these biomass-based ACs produced through physical activation processes still have great advantages over the reference AC. Since, these physical (steam) activation processes are simple, cost-effective, and do not require any chemical activation or post-treatment, which highlights the use of this residues as a source for large scale AC production.

Fig. 5.

Variation of adsorption capacity and adsorbate removal rate at T = 23 °C and equilibrium times t = 60 min for a)180 ppm and b)300 ppm toluene solution using different dosages.

The effects of adsorbent dosage on the adsorption performances of ACs were examined by varying the adsorbent dosage from 10 to 30 mg, see Fig. 5a and b. It was clearly observed that adsorbent mass to liquid volume ratio had influences on the physical interactions among toluene molecules and AC particles. At a low adsorbent dosage, adsorption sites in AC particles were occupied more compared to a high adsorbent dosage. Therefore, an inverse trend was observed between adsorption capacity and adsorbent dosage regardless of the origin of ACs. Whereas the removal rate kept increasing with an increase in adsorbent dosage, however, the removal rate began to show a plateau of around 90% starting from the adsorbent dosage of 20 mg. Thus, based on the above-mentioned experimental outcomes and adsorption process insights, the adsorbent dosage of 20 mg was set as the optimal load for further tests for the adsorption performances of ACs in this work for both pollutant concentrations (180 and 300 ppm).

The adsorption kinetics curves are displayed in Fig. 6a and b where the solid and dashed lines represent the first-order Lagergren and pseudo-second-order kinetic models, respectively. Considering R² values and their relative errors (% εr) for the least square fitting (Table 4) obtained from both models, the pseudo-second-order kinetic model described the adsorption behaviors of toluene onto SWG-based and PT-based AC. At the lower toluene concentration (180 ppm), the adsorption performance of SWG-based AC was better than PT-based AC. In contrast, an opposite trend was observed when the adsorption was performed at the higher toluene concentration (300 ppm) (Fig. 6a and b). It can be inferred that the microporous feature of PT-based AC played an important role in the bulk diffusion onto the surface of PT-based AC. Afterward, intra-particle diffusion of the pollutant into pores of PT-based AC would occur. Interactions with the surface sites by physi- and chemisorption implies a rapid adsorption onto the PT derived active carbon.

Fig. 6.

Variation in the amount of adsorbed toluene at optimal dosage (20 mg/50 mL) with time for the produced and reference activated carbons at a) 180 ppm and b) 300 ppm. The legends indicate the activated carbon source, continuous lines represent the curve fittings for Lagergren model (equation (9)) and dashed lines represents the fitting to pseudo-second order kinetic model (equation (11)).

Table 4.

Relative errors (% εr) for the least square fitting of Lagergren and pseudo-second order kinetic models (PSOK).

Model	Relative errors (% εr) - Toluene adsorption
	AC precursor
	SWG		PT		AC-REF
Sol. Concentration	180 ppm	300 ppm	180 ppm	300 ppm	180 ppm	300 ppm
Lagergren	2.94 (9.1)	3.86 (11.7)	4.71 (11.2)	4.53 (8.4)	2.81 (5.1)	1.64 (4.4)
PSOK	1.61 (3.7)	2.54 (1.1)	1.34 (2.9)	1.24 (0.9)	0.22 (0.8)	0.05 (0.1)

*Value in parenthesis represent the standard deviation calculated from three replicates.

3.7. Isotherms estimations

Three different isotherm models (Langmuir, Freundlich, and Temkin isotherm model) were employed to interpret the adsorption behaviors of toluene onto three different ACs. Considering the R² values obtained from each isotherm model described in Table 5, Langmuir and Freundlich isotherm models (Fig. 7a and b) best described the adsorption behaviors of toluene regardless of the origin of ACs. The R² values of the Langmuir and Freundlich isotherm model (Table 5) indicate that Langmuir isotherm model could depict the existence of an initial monolayer adsorption stage that once achieved is followed by a multilayer adsorption of toluene onto a heterogeneous surface of the biomass-based ACs, which was consistent with the results obtained from BET. According to Treybal [85], the favorability of the adsorption process can be predicted from the n values of the Freundlich isotherm model (Table 5). Values of n in the range 2–10 represent good performance, 1–2 moderately difficult, and less than 1 a poor adsorptive potential. In the case of biomass-based ACs, the magnitude of n (2.0962 for SWG-based AC and 1.9966 for PT-based AC) indicate that the adsorption of toluene onto biomass-based AC performed a good adsorptive process for both precursor materials.

Table 5.

Non-linear isotherm parameters for toluene adsorption onto the produced and reference AC's.

Adsorbent	Langmuir			Freundlich			Temkin
	qm (mg g−1)	KL (L mg−1)	R2	n	KF ((mg g−1)/(mg L−1)1/n)	R2	b (KJ mol−1)	KT (L mg−1)	R2
SWG	2556.8168	0.0090	0.9730	2.0962	116.8108	0.9666	22.0257	7.5752	0.9050
PT	2193.8142	0.0150	0.9767	1.9699	121.7965	0.9730	19.9394	6.6503	0.9023
AC-REF	1751.1794	0.0305	0.9745	1.9936	148.4181	0.9774	18.3008	7.2494	0.9193

Fig. 7.

Non-linear fitting curves: a) Langmuir, b) Freundlich and c) Temkin model for three different AC materials.

In the case of the Temkin isotherm model (Fig. 7c), the R² values reflected acceptable correlation though less than Langmuir and Freundlich (Table 5), respectively. However, the adsorption of toluene arguably does not follow the Temkin isotherm model, which assumes that the adsorption energy of all molecules decreases linearly with the increase in coverage of the adsorbent surface [86].

4. Conclusions

The physical activations of PT and SWG resulted in a combination of micro- and mesoporous, and meso- and macroporous structures, comprising moderate surface areas of 714 and 959 m²/g, respectively. Those parameters are key factors that make them attractive candidates for the removal of aromatic compounds from water. In this study, SWG-based and PT-based AC were employed as an adsorbent and toluene was set as a model pollutant. Adsorption capacities for PT-based and SWG-based AC ranged between 432-716 and 441–711 mg/g, for 180 and 300 ppm of toluene concentration, respectively. The removal rates at the optimal dosage load (20 mg) for the biomass-based ACs were similar to the reference AC (∼90 ± 4%), which confirms the feasibility of biomass-based ACs usage for wastewater treatments under practical conditions. Adsorption kinetics was observed to follow a pseudo-second order model for PT and SWG ACs sample in batch systems both solutions (180 and 300 ppm). The adsorptive performance for the biomass based-ACs followed the Langmuir and Freundlich isotherm models, which refers to the monolayer formation and the later initiation of a multilayer adsorption which is characteristic on heterogeneous adsorbent materials. The combined presence of micropores and mesopores in the AC is thought to be the key on its outstanding kinetic and adsorption performance, in comparison with other commercial porous materials investigated in this study. Moreover, for good development of surface and structural properties as well as textural characteristics, low ash and high contents of C, lignin and fixed carbon biomasses are desired. The use biochar from biomass pyrolysis offers a real opportunity to produce high performance adsorbents towards aromatics removal. Further investigation could be also made to incorporate process modelling tools as well as TEA and LCA valuations to enhance this work.

Eliezer A. Reyes Molina: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Seonghyun Park: Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Sunkyu Park: Contributed reagents, materials, analysis tools or data; Wrote the paper.

Stephen S. Kelley: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Data associated with this study has been deposited at NC State University Libraries Repository, https://www.lib.ncsu.edu/resolver/1840.20/39697.

Appendix A. Supplementary data

The following is the Supplementary data to this article: