Tailored porosity from waste biomass: mesoporous bio-adsorbents for targeted treatment of complex industrial wastewaters

The valorization of waste biomass into tailored adsorbents presents a sustainable strategy for combating industrial water pollution. This study highlights the critical role of precursor morphology in determining the textural properties and function of bio-adsorbents derived from wheat straw (WS) and the organic fraction of municipal solid waste known as biopulp (BP). Through carbonization and KOH activation, the fibrous WS was transformed into a microporous, high-surface-area activated wheat straw (AWS) bio-adsorbent, while the compact BP yielded a mesoporous network in activated biopulp (ABP). This structural difference affects adsorption performance: AWS demonstrated superior efficacy in batch removal of phenols (93.2%) and total organic carbon (85%) from the complex hydrothermal liquefaction aqueous phase (HTL-AP), whereas ABP excelled in treating produced water (PW), achieving >95% removal of organic pollutants. Continuous fixed-bed column studies confirmed the scalability of AWS for HTL-AP treatment, revealing distinct breakthrough dynamics between bulk parameters and specific contaminants. This work provides evidence supporting precursor-dependent tailoring of pore structure for targeted wastewater treatment, providing a possibility for a circular and sustainable solution for the treatment of complex industrial wastewaters.

1 Introduction

The global energy sector faces a dual challenge: meeting energy demands while mitigating the environmental impact of its waste streams. A prime example is produced water (PW), the large volume byproduct from oil and gas extraction. Its composition is highly field-dependent but typically constitutes a complex mixture of dispersed and dissolved hydrocarbons, solids, heavy metals, and inorganic ions. Vast quantities are generated globally; for instance, in 2017 alone, an estimated 24 million tons were produced from the Danish continental shelf (OSPAR Commission, 2019). While physical separation methods like hydrocyclones are commonly used to meet the current legal discharge limit of 30 mg/L for dispersed hydrocarbons (OSPAR Commission, 2001), they are largely ineffective against dissolved polar organic compounds, which are often more bioavailable and environmentally detrimental. The untreated discharge of PW, therefore, presents a persistent risk to aquatic ecosystems.

A similar wastewater challenge emerges from the pursuit of sustainable energy. Hydrothermal liquefaction (HTL) is a promising technology for converting wet biomass into bio-crude oil (Shah et al., 2023). However, for every unit of bio-crude, the process generates a significant fraction of an aqueous phase (HTL-AP) highly concentrated with organic compounds, including carboxylic acids, phenols, and nitrogenous species, as well as nutrients and heavy metals (Utvik, 1999; Seehar et al., 2020; Kohansal et al., 2021; Marrakchi et al., 2023). The high chemical oxygen demand (COD) and toxicity of HTL-AP make its direct disposal environmentally unsustainable, creating a critical bottleneck that hinders the scalability and economic viability of HTL technology. The common thread between these two industrial challenges is the urgent need for cost-effective and sustainable treatment technologies capable of removing dissolved organic pollutants. Adsorption is a particularly attractive method due to its operational simplicity, high efficiency, and potential for resource recovery. The key to its economic feasibility, however, lies in the development of low-cost, high-performance adsorbents (Srivastava et al., 2018). This is where the concept of a circular economy offers a compelling solution: valorizing abundant agricultural and municipal waste streams into functional materials.

Wheat straw (WS), one of the world’s most abundant lignocellulosic agrowaste, represents a vast and underutilized resource. Concurrently, the organic fraction of municipal solid waste (OFMSW), which constitutes 28%–58% of the nearly 2.01 billion tons of MSW generated each year globally, is a significant environmental burden, with production predicted to rise to 3.40 billion metric tons by 2030 (Kohansal et al., 2021). Traditionally viewed as waste, both WS and OFMSW (here processed into “biopulp” or BP) are sustainable sources of carbon that can be transformed into value-added products. Bio-adsorbent, a carbon-rich material produced through the thermochemical conversion of biomass, has emerged as an exceptional bio-adsorbent. Its high porosity, extensive surface area, and tunable surface chemistry enable the effective sequestration of a wide range of pollutants onto its surface, resulting in a cleaned effluent and a nutrient-rich spent adsorbent that could be repurposed in agriculture (Manyuchi et al., 2018). However, the adsorption performance is intrinsically linked to the bio-adsorbent’s physical and chemical properties, which are dictated by the biomass precursor and the activation process.

While previous studies have explored bio-adsorbents for water treatment, a direct comparative study linking the morphological characteristics of distinct waste precursors to their efficacy in treating complex industrial wastewaters like PW and HTL-AP is lacking. Based on this background, this study aims to: (1) synthesize and characterize bio-adsorbents from two morphologically distinct waste streams, fibrous WS and compact BP, through carbonization and chemical activation; (2) evaluate their performance in the batch removal of key pollutants (phenols, COD, TOC, N, P) from HTL-AP and PW under varying conditions; and (3) assess the scalability of the most promising adsorbent in a continuous fixed-bed column system. This research demonstrates a waste-to-wealth strategy, simultaneously addressing the challenges of waste management and industrial wastewater treatment by engineering tailored, low-cost adsorbents for specific pollutant profiles.

2 Materials and methods

2.1 Materials

WS was collected from the wheat crop field located in Aalborg, Denmark. The organic fraction of municipal solid waste, known as BP, was collected from Gemidan Ecogi A/S after undergoing mechanical treatment (Ecogi process). Firstly, the feedstocks were dried at 105 °C for 24 h to obtain dry weight. Later, the dried samples were crushed and sieved to obtain particles with a diameter between 0.25 and 0.5 mm. The ultimate and proximate analysis of the feedstocks is given in Table 1. HTL-AP was obtained from Steeper Energy A/S. It was produced through the hydrothermal liquefaction of mixed waste (biogas digestate + organic waste). PW was received from the Danish Offshore Technology Centre. PW was obtained from the Kraka field (Danish North Sea). Before adsorption experiments, both water phases were centrifuged and filtered to remove oil emulsions and suspended solids.

Table 1

Table 1. Ultimate and proximate analysis of WS and BP (Seehar et al., 2020; Kohansal et al., 2021).

2.2 Bio-adsorbents synthesis

2.2.1 Carbonization of bio-adsorbents

The bio-adsorbent sample was carbonized in a tubular furnace at 800 °C with a heating rate of 10 °C/min, for 90 min (Figure 1). The furnace was first evacuated and then backfilled with argon to remove residual air. This evacuation backfill cycle was performed three times to ensure an inert atmosphere. After establishing the argon environment, a continuous argon purge was maintained throughout the heating stages. The argon flow rate during operation was 3 L/min. After that, the sample was cooled down and rinsed repeatedly with warm distilled water until the filtrate pH reached approximately 6.0–7.0 and dried at 105 °C. The yield (Y) of WS and BP bio-adsorbents was 26.93% and 23.54% respectively, by using the following Equation 1, where Wc and Ws are the weights (g) of the carbonized sample and the original sample used, respectively. m is the moisture content in the sample used.

$Y_{Bio-adsorbent}=\frac{W_c}{Ws-m}\times 100(1)$

Figure 1

Figure 1. Schematic view of biopulp and wheat straw carbonization and activation.

2.2.2 Activation of bio-adsorbents

A two-stage method was used to prepare an activated bio-adsorbent (Figure 1). In the first stage, carbonized bio-adsorbents were chemically activated with KOH. The KOH to carbonized precursor ratio was 1:2 by mass (i.e., 1 g KOH per 2 g carbonized sample) in 200 mL of distilled water. To promote porosity by impregnating potassium ions between the carbon layers, the mixture was stirred for 24 h. Ultrasonication was applied for 1 h to enhance particle dispersion using the instrument VCX 750 (750 W, 20 kHz) with a standard probe (340 g weight, 136 mm length, and 13 mm diameter). The impregnated sample was dried at 60 °C for 24 h to remove water, and then carbonized again in a tubular furnace at 800 °C with a heating rate of 10 °C/min for 90 min. The selected KOH ratio allowed effective activation while considering the limited amount of precursor material.

2.3 Bio-adsorbent characteristics

Nitrogen adsorption–desorption isotherms were recorded using a TriStar III 3000 surface analyzer (Micromeritics Instrument Corp., United States) at 77 K to evaluate the structural evolution of the samples.Specific surface area, pore volume, and pore size of samples were measured with surface characterization analyzer using N₂ at −196 °C. Before analysis, samples were firstly degassed ex situ with N₂ at ambient temperature for 3 hours and then at 125 °C overnight. Subsequently, samples were vacuum degassed in situ at 125 °C for 4 hours. The specific surface area was determined by the Brunauer–Emmett–Teller (BET) method, while the total pore volume and the average pore size were obtained by the Barret-Joyner-Halenda (BJH) desorption analysis (Brunauer et al., 1938; Barrett et al., 1951; Sharma et al., 2023). Microporosity analysis (t-plot, NLDFT/QSDFT) was not performed, as the study focused on mesoporosity and overall surface area, and the amount of material available was limited. The specific surface area was obtained using the BET method, while the pore size distribution was derived from the adsorption branch using the Barrett–Joyner–Halenda (BJH) model. Prior to analysis, all samples were degassed at 300 °C for 12 h under a nitrogen flow.

Surface morphology and elemental composition were examined by scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX). A Tescan Clara UHR SEM was used to acquire SEM micrographs and EDX spectra. The system was operated in analysis mode at 20 keV accelerating voltage and a beam current of 300 pA, with a working distance of approximately 10 mm. SEM images were collected using an Everhart–Thornley (YAG crystal) detector. Elemental mapping was performed using an Oxford Ultim Max 40 mm² EDS detector, and the data were processed with AztecLive Standard software.

2.4 Performance test

2.4.1 Batch adsorption

Before batch adsorption experiments, the pH of the HTL-AP and PW were determined to be 8.7 and 6.6, respectively. The pH of the water phase is an important factor because it affects not only the ionization degree but also the surface properties of the adsorbent (Hoang et al., 2019). Batch adsorption experiments were conducted in a series of Erlenmeyer volumetric flasks, each containing a water phase sample, by adding a specified mass of adsorbent using a thermostatic water bath shaker at a constant speed of 150 rpm (Figure 2). Temperature, contact time, and adsorbent dose significantly impact the adsorption capacity of the adsorption process, which are crucial parameters for optimizing the process at the commercial scale in terms of both energy and economy. In the batch study, HTL-AP and PW were treated using both carbonized (CWS and CBP) and activated (AWS and ABP) adsorbents, synthesized from wheat straw and organic fraction of municipal solid waste (biopulp), respectively. All the materials were used in single dose (1 g/100 mL) and double dose (2 g/100 mL) adsorbent to water phase ratio under both basic and acidic (pH = 4.0) conditions at constant temperature (T = 30 °C) and contact time (24 h) to ensure that adsorption had reached equilibrium. The initial concentration (C_i, mg/L) and residual concentration (C_r, mg/L) in terms of COD, total organic carbon (TOC), total nitrogen (TN), total phosphates (TP), total potassium (TK), and phenols were measured spectrophotometrically before and after adsorption using the LCK cuvette tests placed inside the analysis chamber of UV-visible spectrometer (Hach Lange, DR 3900). Due to the limited amount of sorbent material available, replicate batch experiments could not be performed. Measurements were conducted under carefully controlled conditions to ensure consistency. The removal percentage of adsorbates was calculated by using the following Equation 2.

$Removal\left(\%\right)=\frac{C_i-C_r}{C_i}\times 100(2)$

Figure 2

Figure 2. Schematic view of batch adsorption. Operating conditions: 1–2 g adsorbent to 100 mL contaminated water, 150 rpm stirring, basic and acidic medium, T = 30 °C and 24 h.

2.4.2 Equilibrium adsorption capacity

Equilibrium adsorption capacity ($q_e$, mg g⁻¹) was calculated following Equation 3 as:

$q_e=\frac{\left(C_i-C_r\right)V}{m}(3)$

where $C_i$ and $C_r$ (mg L⁻¹) are the initial and residual concentrations, $V$ is the sample volume (0.1 L), and $m$ is the mass of adsorbent (g; 1 g for 1 g/100 mL and 2 g for 2 g/100 mL). Positive values indicate net uptake; negative values indicate that the measured residual concentration exceeded the initial concentration

2.4.3 Kinetic modeling

Adsorption kinetics were analyzed using pseudo-first order (PFO) and pseudo-second order (PSO) models (Equations 4, 5):

Pseudo-first order (Lagergren) model:

$q_t=q_e\left(1-e^{-k_{1}t}\right)(4)$

Pseudo-second order model:

$q_t=\frac{q_e^2{k}_{2}t}{1+q_e{k}_{2}t}(5)$

where $q_e$ is the equilibrium adsorption capacity (mg/g), and $k_1$ and $k_2$ are the rate constants (1/min or g/mg·min). The experimental $q_t$ values were fitted to these models using nonlinear regression to obtain the kinetic parameters and coefficient of determination ($R^2$) (Table 2).

Table 2

Table 2. Kinetic parameters for TOC, COD, and phenols adsorption (V = 0.1 L, m = 1 g).

2.4.4 Fixed-bed continuous adsorption

For continuous adsorption, AWS and ABP adsorbents were selected for HTL-AP and PW, respectively, due to their high adsorption efficiency during the batch process. Figure 3 shows the experimental setup used for continuous adsorption. The column (H = 50 cm, ID = 1.0 cm) was kept stable by fixing glass beads and a quartz wool layer at the top and the bottom of the column to ensure proper distribution of the inlet solution and to keep a constant mass of adsorbent during the experiment. Before the start of column operation, the bio-adsorbent bed was rinsed with distilled water and left overnight to ensure a compactly packed arrangement of particles without cracks, channels, or voids. The water phase was fed through the fixed-bed column in a down-flow mode at a constant flow rate by an HPLC pump. The flow rate was controlled at 5 mL/min for all experiments with constant bed mass (1.0 g) and depth (3.5 cm). The effluent samples were collected at specified intervals and analyzed for the residual concentration. The experiment was stopped when no further adsorption took place or the column reached exhaustion. The parameters of the adsorption column are given in Table 3. Due to the limited amount of sorbent material available, replicate column experiments could not be performed. Measurements were conducted under carefully controlled conditions to ensure consistency.

Figure 3

Figure 3. Schematic representation of the fixed-bed column setup and lab-scale continuous adsorption setup.

Table 3

Table 3. Parameters of the adsorption column.

3 Results and discussion

3.1 Textural properties and pore structure of bio-adsorbents

The textural properties of the carbonized and activated bio-adsorbents were analyzed by nitrogen adsorption–desorption isotherms and BET analysis, with the results summarized in Table 4.

Table 4

Table 4. Textural properties of the prepared bio-adsorbents.

The carbonized samples exhibited relatively low surface areas and limited pore volumes. CWS showed a BET surface area of 91.53 m²/g with a pore volume of 0.038 cm³/g, while CBP displayed only 37.04 m²/g and 0.051 cm³/g, respectively. Their adsorption isotherms (Figures 4, 5) revealed Type IV and Type III/IV behaviour, characterized by limited adsorption at low relative pressures and a gradual increase at higher pressures, consistent with meso- and macroporous structures. Upon activation, both WS- and BP-derived bio-adsorbents exhibited significant improvements in surface area and porosity. For WS-derived bio-adsorbent, the BET surface area increased from 91.53 m²/g (CWS) to 713.63 m²/g (AWS) upon activation, representing an enhancement of 680%, while the pore volume rose from 0.038 to 0.231 cm³/g (∼6-fold increase). This transformation is associated with the development of abundant mesopores, as indicated by the Type I/IV isotherm and the emergence of a hysteresis loop at higher relative pressures, confirming additional mesopores (Figure 4).

Figure 4

Figure 4. BET surface area analysis N2 adsorption/desorption isotherms CWS and AWS pore-size distribution curve CWS and AWS.

Figure 5

Figure 5. BET surface area analysis N2 adsorption/desorption isotherms CBP and ABP pore-size distribution curve CBP and ABP.

For BP-derived bio-adsorbent, activation also led to notable improvements, though to a lesser extent compared with bio-adsorbents derived from WS. The surface area increased from 37.04 m²/g (CBP) to 158.32 m²/g (ABP), corresponding to a ∼327% increase, while the pore volume rose from 0.051 cm³/g to 0.115 cm³/g (∼125% increase). The resulting Type IV isotherm of ABP revealed a well-developed mesoporous network, which facilitates molecular diffusion and transport.

Overall, activation significantly enhanced the textural properties of both precursors. The AWS bio-adsorbents developed into a mesopore-dominated framework with exceptional surface area, suitable for gas adsorption and catalytic applications, while the ABP bio-adsorbent retained a mesoporous network, making it favourable for diffusion-controlled adsorption, energy storage, and electrochemical applications. These results demonstrate that activation not only amplifies surface area and pore volume but also tailors the pore architecture depending on the biomass precursor.

3.2 Surface morphology of bio-adsorbent

The morphological evolution of the two biomass precursors, WS and BP, during carbonization and activation was systematically examined by SEM (Figure 6). Distinct differences were observed between the two systems, highlighting the critical role of precursor structure in determining the final porous architecture. The raw wheat straw displayed a characteristic lignocellulosic architecture with smooth surfaces, aligned microfibers, and intact vascular bundles (Figure 6a). This ordered fibrous network provided a natural scaffold for further transformation. Upon carbonization, the fibrous framework was preserved but underwent substantial shrinkage and cracking (CWS, Figure 6b). The devolatilization of hemicellulose, cellulose, and lignin generated internal stresses, yielding a more rigid carbon skeleton with rudimentary pores.

Figure 6

Figure 6. SEM micrographs of (a) wheat straw, (b) carbonized wheat straw, (c) activated wheat straw, (d) EDX elements of (b,c,e) raw biopulp, (f) carbonized biopulp (CB), and (g) activated biopulp (AC).

Following chemical activation, the KOH-activated sample (AWS, Figure 6c) exhibited a dramatic textural evolution: the previously coherent surface was extensively etched into an interconnected network of micro- and mesopores with abundant cavities and channels. EDX analysis confirmed the mechanistic role of KOH, with potassium content increasing from 0.4% (CWS) to 2.4% (AWS). The potassium is primarily incorporated into the solid carbon matrix during chemical activation and high-temperature treatment, which minimizes the likelihood of significant potassium leaching into the treated water. Simultaneously, carbon content decreased (91.9%–86.3%) while oxygen content increased (6.3%–10.1%), indicating partial gasification of the carbon framework and incorporation of oxygenated surface functionalities. The etching reactions released gaseous products (H₂, CO₂, CO), which excavated the structure and generated additional porosity. Collectively, these changes converted wheat straw into a mesopore-dominated carbon scaffold with high surface area and tailored chemistry, well-suited for gas adsorption and catalytic applications.

In contrast, raw biopulp presented a fundamentally different microstructure (Figure 6e). Instead of a tubular fibrous network, the surface was compact, dense, and laminated, with sheet-like features indicative of a tightly bound lignocellulosic matrix. This morphology inherently restricts porosity and explains the initially low surface area. Upon carbonization (CBP, Figure 6f), the laminated structure collapsed into a brittle carbon matrix with surface roughness, fractures, and emerging voids caused by the release of volatiles. Although less porous than CWS, this intermediate stage disrupted the dense framework sufficiently to allow chemical agents to penetrate. The most profound transformation occurred after activation (ABP, Figure 6g), where KOH etching generated abundant pores, cavities, and etched tunnels across the fractured surface. This conversion into a mesoporous scaffold provided extensive active sites and enhanced accessibility for adsorbates. EDX analysis confirmed potassium incorporation and a concomitant decrease in carbon content alongside oxygen enrichment, reflecting similar activation mechanisms to those observed in AWS, though with precursor-dependent outcomes. As with AWS, the incorporated potassium in ABP is primarily bound within the carbon matrix, which reduces the likelihood of leaching into treated water and supports the sustainable treatment claim.

The SEM and EDX analyses reveal that precursor morphology dictates the activation pathway and final pore architecture. WS, with its inherently fibrous and open framework, facilitated deep penetration of KOH, producing a mesopore-rich network with exceptionally high surface area. By contrast, BP’s compact laminated morphology initially resisted pore formation, but ultimately yielded a mesopore-dominated structure upon activation. This distinction indicates a key materials-design principle: fibrous precursors favor microporosity and high surface area, while compact precursors generate mesoporous networks more suited for diffusion-controlled processes such as liquid-phase adsorption and electrochemical storage. In both cases, the combined action of thermal decomposition and KOH etching proved highly effective, transforming biomass into functional porous carbons. The general mechanism involves sequential solid–solid and solid–liquid reactions between KOH and carbon, accompanied by the reduction of K⁺ ions to metallic K, oxidation of carbon to oxides and carbonates, and evolution of gaseous products. These processes collectively restructure the carbon matrix, creating hierarchical porosity while simultaneously introducing surface functionalities that enhance adsorption and reactivity.

3.3 Effect of bio-adsorbents on batch adsorption

The efficacy of the synthesized bio-adsorbents, CWS, AWS, CBP, and ABP, was evaluated for the remediation of two challenging aqueous waste streams: HTL-AP and PW. The performance was assessed against a suite of critical water quality parameters, including water-soluble phenols, total TOC, COD, TN, TP, and TK. The results, detailed in Figures 7, 8, demonstrate a compelling structure-activity relationship, where the adsorption performance is intrinsically governed by the textural and chemical properties of the adsorbents, as established in Section 3.1 and Section 3.2, in conjunction with the complex chemistry of the target aqueous matrix. Since HTL-AP and PW possess inherent, non-adjustable compositions, the present study evaluated adsorption performance at their native pollutant levels. Therefore, adsorption isotherms based on multiple initial concentrations could not be established. Due to the limited amount of adsorbent produced and the large number of measurements required, replicate experiments could not be performed. Nevertheless, all measurements were conducted under carefully controlled conditions to maintain consistency and comparability across experiments. Instead, the equilibrium adsorption capacity (qe) is reported for phenols, TOC, and COD as a direct measure of the adsorbent’s performance under realistic conditions. Future work will involve controlled multi-concentration isotherm experiments with model solutions to determine q_max and fit Langmuir/Freundlich models.

Figure 7

Figure 7. Concentration of various contaminants before and after batch adsorption of HTL-AP (a) and PW (b).

Figure 8

Figure 8. Percentage removal of Phenols, COD, TOC, TN, TK, and TP after batch adsorption.

3.3.1 Phenol removal: a function of porosity and solution pH

Phenol adsorption exhibited significant dependence on the adsorbent’s pore architecture and the solution pH. The HTL-AP, characterized by a high pH of 8.7 and complex composition, presented a challenging environment for adsorption. As shown in Figure 8, the AWS bio-adsorbent, endowed with an exceptionally high BET surface area (713.63 m²/g) and a mesopore-dominated network (Section 3.1), demonstrated the highest affinity for phenols. A single dose of AWS achieved a notable phenol removal of 79.54% from HTL-AP. This performance was further enhanced to 93.20% and 86.00% with a double dose of AWS under basic and acidic pH conditions, respectively. This superior performance is attributed to the vast microporous surface area of AWS, which provides abundant adsorption sites for the relatively small phenol molecules.

In contrast, the ABP, with a more mesoporous structure (BET surface area: 158.32 m²/g), showed lower efficacy for phenol removal from HTL-AP, with a maximum of 67.68% achieved using a double dose. This aligns with the expectation that microporous materials are more effective for the adsorption of small organic molecules like phenol. The critical role of solution chemistry was further underscored by the adsorption results from PW, which has a milder, near-neutral pH (6.6). In this environment, the phenol removal efficiencies were markedly higher across all adsorbents. Strikingly, a single dose of ABP achieved a near-complete phenol removal of 99.16% from PW. This profound pH dependence is well-documented; the ionization state of the phenolic compound and the surface charge of the carbon adsorbent are modulated by pH, favoring adsorption under neutral to acidic conditions, where the phenol is predominantly in its neutral, less hydrophilic form (Singh et al., 2008; De La Luz-Asunción et al., 2015).

The equilibrium adsorption capacities for phenols (Tables 5, 6) show that AWS is the most effective sorbent for phenolic removal from HTL-AP, reaching a maximum $q_e=4.12$ mg g⁻¹ at a 1 g dose. Increasing the dose to 2 g reduced the capacity (to $q_e=2.41$ mg g⁻¹), consistent with typical surface-saturation and dilution effects. Acidification (pH 4) did not significantly enhance phenol uptake for AWS, suggesting that phenolic adsorption is not strongly pH-driven under the studied conditions. In PW, all sorbents displayed low phenol capacities ($q_e\approx 0.05$–0.11 mg g⁻¹), which reflects the very low native phenol concentration (1.18 mg L⁻¹) and the limited driving force for adsorption. Overall, phenol removal across both streams is modest and adsorbent-dependent, with AWS showing the most consistent performance.

Table 5

Table 5. Equilibrium adsorption capacities for HTL-AP (native concentrations).

Table 6

Table 6. Equilibrium adsorption capacities for Produced Water (PW, native concentrations).

3.3.2 Removal of broad-spectrum organic pollutants (TOC and COD)

The removal of bulk organic load, quantified as TOC and COD, further highlighted the complementary roles of the different adsorbents. The HTL-AP, with an extremely high initial TOC of 18.98 mg/L and COD of 45.45 mg/L, represents a highly complex matrix containing diverse organic functional groups (Silva Thomsen et al., 2022). For such a challenging stream, the high-surface area, microporous AWS bio-adsorbents were again the most effective. As depicted in Figure 8, a double dose of AWS under acidic conditions achieved a TOC reduction of 85% and a COD reduction of 65.56%, the highest among all treatments for HTL-AP. This suggests that the extensive surface area of AWS is crucial for adsorbing the wide variety of organic compounds present.

Conversely, for the less contaminated PW (TOC: 1.86 mg/L), the mesoporous structure of ABP proved highly effective, likely facilitating better diffusion and adsorption of larger organic molecules or aggregates. A single dose of ABP under basic conditions removed 95.94% of TOC, and under acidic conditions, it removed 92.25% (Figure 8). Interestingly, a double dose of ABP reduced this efficiency to 52.04%, potentially indicating competitive adsorption or pore blocking effects in a less saturated system. The COD removal trends mirrored those of TOC, with ABP achieving a remarkable 94.99% COD removal from PW under acidic conditions with a single dose.

The adsorption trends for nitrogen and phosphorus followed a similar pattern to the organic pollutants, though with varying efficiencies. The maximum TN removal from HTL-AP was 54.85%, achieved with a double dose of AWS under acidic conditions. For PW, a single dose of ABP under acidic conditions removed 82.59% of TN. The incorporation of oxygenated functionalities during activation (as indicated by the findings of the EDX measurements in Section 3.2) likely enhanced specific interactions with nitrogenous compounds through polar and hydrogen bonding. Notable phosphorus removal was observed only in HTL-AP, which contained a significant concentration of TP (0.20 g/L). AWS was the most effective, removing 73.97% of TP under acidic conditions with a double dose. The negligible TP content in PW resulted in no significant removal. In contrast, no adsorbent showed significant efficacy for potassium (TK) removal from either stream, as the ionic potassium likely competes poorly with organic molecules for adsorption sites and may itself be released from the KOH-activated bio-adsorbents.

The batch adsorption results unequivocally demonstrate that there is no universal “best” adsorbent; rather, optimal performance is dictated by a synergy between the adsorbent’s physical/chemical properties and the wastewater’s composition. The microporous, high-surface-area AWS is from the tested materials, the material of choice for treating complex, high-strength wastewaters like HTL-AP, where maximizing the adsorption capacity for a diverse array of small organic molecules (phenols, TOC) is paramount.

The mesoporous ABP excels in treating the less complex stream PW, where its pore network facilitates superior diffusion and removal of bulk organics (TOC, COD) under favorable pH conditions. This structure-function relationship, established through meticulous textural and morphological analysis, provides a powerful blueprint for the rational design of targeted adsorption systems for sustainable wastewater remediation. The ability to tailor biomass-derived carbons for specific aqueous waste challenges marks a significant advancement towards circular economy principles in water treatment.

Adsorption behaviour for bulk organics (TOC and COD) was markedly different from phenols. In HTL-AP, AWS exhibited substantial capacity for COD ($q_e=460$ mg g⁻¹ at 1 g), and this increased dramatically under acidic conditions ($q_e=1490$ mg g⁻¹ at pH 4), indicating that protonation of functional groups and increased hydrophobic interactions strongly favour adsorption of complex organic matrices. Likewise, TOC uptake increased from moderate levels at neutral pH (24–57 mg g⁻¹) to very high capacity at pH 4 ($q_e=815$ mg g⁻¹). In PW, ABP demonstrated the highest TOC capacities, particularly under acidic conditions (up to $q_e=178.84$ mg g⁻¹), while several sorbents showed minimal or negative COD capacities, reflecting either analytical variability or desorption/leaching phenomena. Overall, adsorption of broad-spectrum organics is highly responsive to pH and adsorbent chemistry, with activated bio-adsorbents showing strong potential for bulk organic reduction in both HTL-AP and PW.

3.4 Continuous flow adsorption and breakthrough behaviour

An assessment of adsorbent performance under continuous-flow conditions is imperative for translating batch-scale efficacy to practical, industrial applications. The breakthrough curve is the fundamental tool for evaluating the dynamic performance of a fixed-bed adsorption column, defining its separation efficiency and operational lifespan. It describes the effluent concentration profile over time, expressed as the ratio Ct/C0, where Ct is the effluent concentration (mg/L) and C0 is the initial adsorbate concentration (mg/L). The shape of this curve, determined by the breakthrough time (tb), the exhaustion time (t_e), and the slope of the mass transfer zone (MTZ), provides critical parameters for the design and scale-up of adsorption systems, including bed height, flow rate, and regeneration cycles. Given its exceptional performance in batch tests for the HTL-AP (c.f., Section 3.3), AWS was selected for continuous-flow column experiments. Breakthrough curves were determined for the key wastewater parameters, TOC, COD, and phenols. Experiments were conducted under constant conditions of a 5 mL/min flow rate and a bed height of 3.5 cm; the resulting profiles are presented in Figure 9.

Figure 9

Figure 9. Breakthrough curve of TOC, COD, and Phenols adsorption on AWS using HTL-AP.

The breakthrough curves for TOC and COD exhibit strikingly similar trajectories, both characterized by a sharp, sigmoidal shape. This indicates a favorable adsorption process with a relatively narrow and stable mass transfer zone propagating through the column bed. The near-synchronous saturation and almost identical exhaustion times for TOC and COD suggest that the diverse suite of organic compounds constituting these bulk parameters in HTL-AP are adsorbed by AWS through comparable mechanisms and at similar rates. Notably, the breakthrough for both TOC and COD occurred significantly faster than for phenols. This earlier exhaustion is a direct consequence of the competitive adsorption landscape within the complex HTL-AP matrix. The rapid saturation of adsorption sites by the high concentration of various organic compounds (e.g., carboxylic acids, ketones, esters) increases the rate of mass transfer into the bed, effectively widening the MTZ and leading to a premature breakthrough of the bulk organic load.

In contrast, the breakthrough curve for phenols is more gradual, exhibiting a shallower slope and a delayed exhaustion time. This indicates a stronger affinity and specific adsorption mechanism for phenolic compounds onto the microporous AWS surface. The delayed breakthrough suggests that phenols successfully compete for high-energy sites within the mesopore (c.f., Section 3.1), or that their adsorption kinetics are governed by a slower, more selective diffusion process. This finding robustly corroborates the batch adsorption results, which demonstrated AWS’s high-capacity removal of phenols. Separate continuous-flow experiments were conducted with PW. While the column operation was hydraulically stable, the removal efficiency for all measured parameters was negligible. This lack of efficacy is attributed to the specific history of the PW sample, which was approximately 6 years old. Extended storage likely led to significant alterations in wastewater composition, including volatilization of light organic fractions, oxidative degradation, and potential microbial activity. These processes undoubtedly altered the adsorbability of the constituents, rendering them less amenable to removal by adsorption onto AWS, and highlighting the importance of using representative, fresh samples for treatability studies.

The continuous-flow data confirms that AWS is a viable adsorbent for mitigating the organic load of HTL-AP in a fixed-bed system. The differential breakthrough behavior reveals a key design consideration: a larger bed volume or a series of columns would be required to achieve significant TOC/COD reduction before the bed is exhausted. However, the column exhibits a strong capacity for phenolic compounds, indicating its potential as a polishing step for targeted removal of these more recalcitrant and toxic contaminants after preliminary treatment. Future work will focus on optimizing hydraulic loading conditions, determining the kinetics of the mass transfer zone, and evaluating the regeneration potential of the spent AWS adsorbent to assess economic feasibility.

4 Conclusion

This study establishes a link between biomass precursor morphology and the function of resulting bio-adsorbents for wastewater remediation. We demonstrate that the fibrous architecture of wheat straw directs KOH activation toward a microporous, high-surface-area carbon (713.63 m²/g), while the compact structure of biopulp before KOH activation yields a mesoporous network (158.32 m²/g). This structural determinism impacts the performance: AWS shows good performance in the targeted adsorption of phenols and organics from complex, high-strength HTL-AP, while ABP has potential for the rapid removal of bulk pollutants from PW. Continuous-flow column studies confirm the efficacy of AWS for HTL-AP treatment and reveal distinct breakthrough dynamics for bulk parameters versus specific contaminants, providing important design insights. By transforming waste biomass into tailored adsorbents, this work introduces a possibility for a sustainable, circular strategy for addressing the challenges of complex industrial wastewater.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

Author contributions

ST: Conceptualization, Funding acquisition, Investigation, Methodology, Writing – original draft, Writing – review and editing. KS: Data curation, Formal Analysis, Writing – review and editing. RK: Formal Analysis, Writing – review and editing. MM: Formal Analysis, Writing – review and editing. AA: Data curation, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This study was performed within the Radical Innovation Sprint project (RIS-2022) funded by the Danish Offshore Technology Centre. RK acknowledges financial support from the Danish National Research Foundation (grant no. DNRF189) through the Center for Sustainable Energy Materials (CENSEMAT). The Carlsberg Foundation (Grant no: CF20-0364) and Aarhus University Centre for Integrated Materials Research (iMat) are acknowledged for funding the Tescan Clara SEM.

Acknowledgements

The authors gratefully acknowledge the Danish Offshore Technology Centre, Denmark, and the Department of Materials and Production, Aalborg University, Aalborg, for their experimental support during the carbonization of the feedstocks.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Barrett, E. P., Joyner, L. G., and Halenda, P. P. (1951). The determination of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms. J. Am. Chem. Soc. 73, 373–380. doi:10.1021/ja01145a126

CrossRef Full Text | Google Scholar

Brunauer, S., Emmett, P. H., and Teller, E. (1938). Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 60, 309–319. doi:10.1021/ja01269a023

CrossRef Full Text | Google Scholar

De La Luz-Asunción, M., Sánchez-Mendieta, V., Martínez-Hernández, A. L., Castaño, V. M., and Velasco-Santos, C. (2015). Adsorption of phenol from aqueous solutions by carbon nanomaterials of one and two dimensions: kinetic and equilibrium studies. J. Nanomater. 2015, 405036. doi:10.1155/2015/405036

CrossRef Full Text | Google Scholar

Hoang, L. P., Van, H. T., Nguyen, L. H., Mac, D. H., Vu, T. T., Ha, L. T., et al. (2019). Removal of Cr (vi) from aqueous solution using magnetic modified biochar derived from raw corncob. New J. Chem. 43 (47), 18663–18672. doi:10.1039/c9nj02661d

CrossRef Full Text | Google Scholar

Kohansal, K., Sharma, K., Sohail Toor, S., Lozano Sanchez, E., Zimmermann, J., Rosendahl, L. A., et al. (2021). Bio-crude production improvement during hydrothermal liquefaction of biopulp by simultaneous application of alkali catalysts and aqueous phase recirculation. Energies 14, 4492. doi:10.3390/en14154492

CrossRef Full Text | Google Scholar

Manyuchi, M. M., Mbohwa, C., and Muzenda, E. (2018). Potential to use municipal waste bio char in wastewater treatment for nutrients recovery. Phys. Chem. Earth 107, 92–95. doi:10.1016/j.pce.2018.07.002

CrossRef Full Text | Google Scholar

Marrakchi, F., Toor, S. S., Nielsen, A. H., Pedersen, T. H., and Rosendahl, L. A. (2023). Bio-crude oils production from wheat stem under subcritical water conditions and batch adsorption of post-hydrothermal liquefaction aqueous phase onto activated hydrochars. Chem. Eng. J. 452, 139293. doi:10.1016/j.cej.2022.139293

CrossRef Full Text | Google Scholar

OSPAR Commission (2001). OSPAR recommendation 2001/1 for the management of produced water from offshore installations.

Google Scholar

OSPAR Commission (2019). Denmark assessment of discharges, spills and emissions from offshore oil and gas installations in 2013.

Google Scholar

Seehar, T. H., Toor, S. S., Shah, A. A., Pedersen, T. H., and Rosendahl, L. A. (2020). Biocrude production from wheat straw at sub and supercritical hydrothermal liquefaction. Energies 13 (12), 3114. doi:10.3390/en13123114

CrossRef Full Text | Google Scholar

Shah, A. A., Sharma, K., Seehar, T. H., Toor, S. S., Sandquist, J., Inge, S., et al. (2023). Sub-supercritical hydrothermal liquefaction of lignocellulose and protein-containing biomass. Fuels 5, 75–89. doi:10.3390/fuels5010005

CrossRef Full Text | Google Scholar

Sharma, K., Kohansal, K., Azuara, A. J., Rosendhal, L. A., Benedetti, V., Yu, D., et al. (2023). Green and facile recycling of bauxite residue to biochar-supported iron-based composite material for hydrothermal liquefaction of municipal solid waste. Waste Manag. 71, 259–270. doi:10.1016/j.wasman.2023.08.024

PubMed Abstract | CrossRef Full Text | Google Scholar

Silva Thomsen, L. B., Anastasakis, K., and Biller, P. (2022). Wet oxidation of aqueous phase from hydrothermal liquefaction of sewage sludge. Water Res. 209, 117863. doi:10.1016/j.watres.2021.117863

PubMed Abstract | CrossRef Full Text | Google Scholar

Singh, K. P., Malik, A., Sinha, S., and Ojha, P. (2008). Liquid-phase adsorption of phenols using activated carbons derived from agricultural waste material. J. Hazard. Mater. 150 (3), 626–641. doi:10.1016/j.jhazmat.2007.05.017

PubMed Abstract | CrossRef Full Text | Google Scholar

Srivastava, S., Agrawal, S. B., and Mondal, M. K. (2018). Fixed bed column adsorption of Cr(VI) from aqueous solution using nanosorbents derived from magnetite impregnated Phaseolus vulgaris husk. Environ. Prog. Sustain. Energy 38 (S1), 68–76. doi:10.1002/ep.12918

CrossRef Full Text | Google Scholar

Utvik, T. I. R. (1999). Chemical characterisation of produced water from four offshore oil production platforms in the North Sea. Chemosphere 39, 2593–2606. doi:10.1016/S0045-6535(99)00171-X

CrossRef Full Text | Google Scholar