Combined Continuous Resin Adsorption and Anaerobic Digestion of Olive Mill Wastewater for Polyphenol and Energy Recovery

Abstract

Olive mill wastewater (OMWW) has high energetic potential due to its organic load, but its complex composition and toxicity limit efficient energy recovery. This study proposes an innovative integrated process combining continuous resin adsorption with anaerobic digestion to detoxify OMWW and recover renewable energy simultaneously. It studies the recovery of polyphenols, methane production, and substrate degradation efficiency using resin column bed heights (C1: 5.7 cm, C2: 12.1 cm, C3: 18.5 cm), as well as kinetic modeling of organic matter degradation. Adsorption reduced chemical oxygen demand (COD) by up to 80% and polyphenols by up to 64%, which significantly improved substrate biodegradability from 34% to 82%, corresponding to a methane yield of 287 mL CH₄/g COD. Organic matter was fractioned into rapid (S₁), moderate (S₂), and slow (S₃) biodegradable fractions. The highest degradation kinetics was C3, with methane production rates of K₁ = 23.86, K₂ = 2.47, and K₃ = 2.92 mL CH₄/d. However, this condition produced the lowest volumetric methane production due to excessive COD removal, including readily biodegradable matter. These results highlight the importance of optimizing the adsorption step in order to find to a balance between detoxification and energy recovery from OMWW, thus supporting the principles of circular economy and promoting renewable energy production.

1. Introduction

More than 95% of olive oil is produced in the Mediterranean region. During the 2023–2024 season, global olive oil production was estimated at approximately 3.1 million tons, which represents about 9 to 11 million cubic meters of OMWW [1,2]. The global annual production of OMWW is estimated at between 4.5 and 7.5 million cubic meters [3]. Inadequate treatment of OMWW, such as land application or direct discharge into aquatic systems, results in significant environmental problems, including soil acidification and groundwater contamination [4,5,6,7]. In 2023, global olive oil production exceeded 3.2 million tons, generating more than 10 million m³ of OMWW. This effluent is characterized by high organic content, low pH, and high concentrations of toxic phenolic compounds, making its treatment both urgent and technically challenging [8,9].

Various treatment methods, such as aerobic processes, advanced oxidation, and membrane filtration, have been investigated. Recent studies have focused on hybrid systems to overcome phenolic inhibition in wastewater treatment. Li et al. [10] showed that adsorption combined with electro-Fenton process improved phenol degradation, while Agabo-García et al. [11] developed a multistage approach combining advanced oxidation, flocculation, and filtration for OMWW detoxification. However, these processes are often associated with high operational costs, high energy requirements, or potential production of secondary pollutants [12,13,14,15,16]. In addition, anaerobic digestion (AD) is a promising approach for OMWW treatment, with the dual benefit of organic matter degradation and renewable energy recovery in methane-rich biogas [17,18].

OMWW is classified as one of the most concentrated liquid agro-industrial wastes due to its extremely high organic content [19], and has significant energy recovery potential [20]. However, its intrinsic properties include low alkalinity, a deficiency in ammonium nitrogen, a high total organic carbon content, an acidic pH (4.8–5.7) [21], and an extremely high chemical oxygen demand (COD) of up to 220 g/L [22]. The reported biological oxygen demand (BOD5, 35–110 g/L) [17] is high but still low compared to the very high COD values, reflecting the typical low biodegradability of these effluents.

The high concentration of polyphenolic compounds (1–10 g/L) in OMWW is a major barrier to conventional anaerobic digestion systems [23,24,25]. Various studies have shown that the inhibitory effects of polyphenols on AD are concentration-dependent. Calabrò et al. [26] investigated the impact of different concentrations of polyphenols (0.5, 1, and 2 g/L) on methane production. The study found that these compounds mainly affect methanogenesis, leading to accumulation of volatile fatty acids (VFAs) due to the continued activity of acidogenic bacteria while methanogens are inhibited. VFAs disrupt methanogenic archaea by inhibiting key enzymes and damaging cell membranes, resulting in process instability and reduced methane production [9,27,28,29,30,31].

To overcome these challenges and stabilize anaerobic digestion, effective pH management and removal of polyphenolic compounds are crucial to optimize the efficiency of anaerobic digestion [32]. Despite their inhibitory effects, polyphenols have valuable properties, including antioxidant, anti-cancer, antiviral, and anti-inflammatory activities [9]. These properties make them highly attractive for applications in the pharmaceutical and food industries, aligning with the principles of circular economy and offering opportunities for environmental protection and economic gain [33].

Several studies have investigated the anaerobic digestion of raw OMWW, demonstrating its potential for biogas production but also highlighting significant limitations. Khoufi et al. [34] reported that methane production was significantly inhibited when the reactor was fed with raw OMWW at a loading rate of 2–4 g COD/L/d, resulting in a decrease in methane yield from an average of 0.15 L CH₄/g COD to 0.05 L CH₄/g COD. In addition, Erraji et al. [35] showed total inhibition of batch anaerobic digestion of OMWW. However, some recent studies have shown that the incorporation of activated carbon and iron improves anaerobic digestion of wastewaters containing phenolic compounds [36]. Madigou et al. [37] proposed an acclimatation strategy to increase the phenol tolerance of anaerobic digestion consortia. This can be supported by the results of a full-scale anaerobic digestion plant treating olive mill by-products made of olive pulp and pitted pomace. Although polyphenol concentrations ranged from 1.8 to 3 g equivalent gallic acid per kg, methane production of 700 L/kgVS and a reduction in polyphenol concentration were obtained [38].

Another strategy is the removal of phenolic compounds, which was shown to improve methane yield and degradation efficiency [39]. Levén et al. [40] observed a negative correlation between phenolic content and microbial performance. Adsorption technology is a cost-effective and sustainable solution for OMWW treatment, characterized by its operational simplicity and the ability to regenerate sorbents for repeated use [41]. Several studies have highlighted the effectiveness of XAD-4 resin for the adsorption and recovery of OMWW polyphenols [30]. The integration of anaerobic digestion with adsorption offers an innovative approach to OMWW valorization. This combined strategy mitigates the inhibitory impact of polyphenols while allowing the recovery of bioactive compounds with potential applications in the pharmaceutical and food industries. Although resin adsorption and anaerobic digestion have been studied individually, their integrated application for OMWW is still relatively under-explored. Notably, there is a lack of comprehensive kinetic models that can accurately describe and predict methane production dynamics in such integrated systems. A deeper understanding of methane production kinetics and substrate degradation pathways is essential to improve the efficiency and optimization of anaerobic digestion processes [42]. Kinetic modeling allows the determination of the different organic matter fractions in OMWW based on their rates of biodegradation, leading to a thorough understanding of their behavior, including rapid, moderate, and slow kinetics.

This study addresses the following objectives by integrating continuous resin adsorption with batch anaerobic digestion: assessing the efficiency of Amberlite XAD-4 resin in removing polyphenolic and organic compounds from OMWW and evaluating the effectiveness of anaerobic digestion in conjunction with resin pretreatment and create and implement a kinetic model to predict methane production during raw and pretreated olive mill effluent anaerobic digestion.

2. Materials and Methods

2.1. Materials

OMWW was collected from a traditional Moroccan olive oil station in Oujda and stored at 4 °C to preserve its characteristics before use in experiments. Before adsorption, the OMWW was filtered using filter paper to remove suspended solids and prevent clogging of the column. Amberlite macroporous XAD-4 resin was used as an adsorbent to extract polyphenolic compounds from OMWW. The main technical characteristics of the resin are detailed in

Table 1. Technical characteristics of Amberlite XAD-4 resin.

Parameter	Surface Area (m2/g)	Pore Diameter (nm)	Mean Pore Size (Å)	Particle Size (mm)
XAD-4 resin	725	15	100	0.46–0.69

. The XAD-4 resin, acetone, gallic acid, Folin–Ciocalteu reagent, sodium carbonate, and ethanol were purchased from Sigma-Aldrich (Burlington, MA, USA). Before use, the resin was soaked in acetone for 8 h, followed by two rinses with deionized water and air-drying at room temperature.

The inoculum used for anaerobic digestion tests was an anaerobic sludge obtained from a mesophilic UASB reactor (Upflow Anaerobic Sludge Blanket) operating on wastewater from the Saica paper factory (Laveyron, France). After reception, the inoculum was stored and stirred under mesophilic conditions (35 °C) and was periodically fed with ethanol in order to maintain methanogenic activity. Before starting the tests, the inoculum was subjected to a 5-day starvation phase to remove residual organic matter. Sodium bicarbonate, at a concentration of 2.5 g/L, was then added to buffer the system to maintain optimal pH levels for anaerobic digestion.

2.2. Analysis

The OMWW effluent pH was measured using a Hanna HI 5221 pH meter. The total solid (TS) and volatile solid (VS) concentrations were carried out following the APHA (American Public Health Association) method. Total polyphenolic compounds (TPhC) were quantified using the Folin–Ciocalteu colorimetric technique. A total of 20 µL of a diluted sample was combined with 500 µL of diluted Folin–Ciocalteu reagent, followed by 400 µL of 7.5% (w/w) sodium carbonate solution. The solution was incubated at 35 °C for 20 min and thereafter analyzed in triplicate using a Shimadzu UV-1800 spectrophotometer (Shimadzu Corporation, Kyoto, Japan) at 760 nm. Results were expressed in milligrams of gallic acid equivalent per liter [43]. The total chemical oxygen demand (TCOD) concentration was determined in raw OMWW, filtered OMWW and resin-pretreated samples. In this analysis, 0.5 g of each sample was hydrolyzed in 20 mL of sulfuric acid and allowed to digest overnight to ensure complete hydrolysis. The hydrolysis process allowed for accurate sample dilution and facilitated the complete conversion of organic matter into CO₂ and water, allowing for precise measurements of TCOD. The TCOD was then measured spectrophotometrically utilizing Aqualytic COD Vario tubes (0–1500 mg O₂/L). For each analysis, 2 mL of a diluted hydrolyzed sample was transferred to Spectroquant test kits in triplicate and 2 mL of Milli-Q water (Merck, Darmstadt, Germany) was used as a blank. The tubes were heated in a HACH COD reactor at 150 °C for 2 h to oxidize the organic matter using potassium dichromate (K₂Cr₂O₇) in an acidic environment. They were then cooled to ambient temperature for 1 h. COD measurements were performed using a HACH DR/2000 spectrophotometer (Hach Company, Loveland, CO, USA) at 620 nm wavelength [44].

2.3. Continuous Adsorption Experiments

Continuous adsorption tests were performed in a fixed-packed bed configuration using a glass column of 1.6 cm in diameter and 30 cm in height. The ends of the column were filled with cotton wool to ensure a constant effluent flow. The OMWW solution, filtered using a filter paper and diluted to 50% with distilled water, was introduced to the XAD-4 resin column in an up-flow configuration through a peristaltic pump. The studies were conducted at a constant flow rate of 2 mL/min in three different bed heights (Z = 5.7, 12.1, and 18.5 cm), corresponding to 2.5, 5, and 10 g of dry adsorbent, respectively. The flow rate of 2 mL/min was selected based on preliminary tests to ensure a sufficient contact time without clogging. This value also aligns with previous studies on phenolic wastewater adsorption in fixed-bed columns [45,46]. As polyphenol concentration was measured offline, adsorption tests lasted longer than the time necessary to reach saturation of the column.

2.4. Column Performance

The ratio of outlet concentration (C_t) to inlet concentration (C₀) serves as an indicator of the retention capacity of the adsorbent for polyphenolic compounds. The performance of a fixed-bed column can be assessed by breakthrough curves, which plot the C_t/C₀ ratio over time. The main characteristics obtained from these breakthrough curves provide indications of the adsorption efficiency and operational performance of the fixed-bed system [40,41]. When the bed reaches full saturation, i.e., when the outlet concentration is equal to the inlet concentration and remains constant thereafter, the total adsorption capacity of the column (q_total, mg) was calculated. It represents the amount of polyphenolic compounds retained by the resin fixed bed. This value is obtained by integrating the area under the breakthrough curve [42], as shown in Equation (1):

$${\mathrm{q}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}={\displaystyle \frac{\mathrm{Q}{\mathrm{C}}_0}{1000}}{\int }_{\mathrm{t}=0}^{\mathrm{t}={\mathrm{t}}_{\mathrm{s}}}\left(1-{\displaystyle \frac{{\mathrm{C}}_{\mathrm{t}}}{{\mathrm{C}}_0}}\right)\mathrm{d}\mathrm{t}$$

(1)

where Q is the column feed flow rate (L/min), C_t is the outlet concentration (mg/L), C₀ is the inlet concentration (mg/L), and t_s is the time required for the bed to become saturated (min).

The resin adsorption capacity (${\mathrm{q}}_{\mathrm{e}}$, mg/g) is determined by the amount of polyphenol recovered by the fixed bed per gram of adsorbent (m resin) present in the bed and is calculated from Equation (2):

$${\mathrm{q}}_{\mathrm{e}}={\displaystyle \frac{{\mathrm{q}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}{\mathrm{m} \mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{n}}}$$

(2)

The polyphenol adsorption yield (%) was calculated based on the cumulative amount of polyphenols adsorbed until column saturation, derived from the area under the breakthrough curve. The yield was calculated as the ratio of ${\mathrm{q}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ to the total mass of phenolics introduced to the column (m_t) using the following expression (3)

$$\mathrm{P}\mathrm{o}\mathrm{l}\mathrm{y}\mathrm{p}\mathrm{h}\mathrm{e}\mathrm{n}\mathrm{o}\mathrm{l} \mathrm{a}\mathrm{d}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}\left(\mathrm{\%}\right)={\displaystyle \frac{q_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}{m_{\mathrm{T}\mathrm{P}\mathrm{h}\mathrm{C} \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}}$$

(3)

where ${\mathrm{q}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ is the total amount of polyphenols adsorbed (mg), obtained through breakthrough curve integration as described in Equation (1), and ${\mathrm{m}}_{\mathrm{T}\mathrm{P}\mathrm{h}\mathrm{C} \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ is the total mass of polyphenols introduced into the column until saturation (mg).

The TCOD removal yield (Y_{removal, TCOD}) was determined using Equation (4):

$${\mathrm{Y}}_{\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l}, \mathrm{T}\mathrm{C}\mathrm{O}\mathrm{D}} \left(\mathrm{\%}\right)={\displaystyle \frac{\mathrm{T}{\mathrm{C}\mathrm{O}\mathrm{D}}_{\mathrm{i}}-{\mathrm{T}\mathrm{C}\mathrm{O}\mathrm{D}}_{\mathrm{f}}}{\mathrm{T}{\mathrm{C}\mathrm{O}\mathrm{D}}_{\mathrm{i}}}}\times 100$$

(4)

where $\mathrm{T}{\mathrm{C}\mathrm{O}\mathrm{D}}_{\mathrm{i}}$ and TCOD_f represent the initial and final total COD concentrations, respectively.

In addition to capacity calculations, key hydrodynamic parameters, including the empty-bed contact time (EBCT), superficial velocity, and Reynolds number were calculated to confirm the flow regime and ensure consistency between tests. The EBCT was determined using the ratio of empty bed volume to flow rate, while the superficial velocity was calculated from the flow rate and column cross-sectional area.

2.5. Biochemical Methane Potential (BMP) Assays

Batch BMP tests were performed in triplicate on raw and pretreated OMWW to evaluate the effect of adsorption pretreatment on anaerobic digestion kinetics and methane potential. The BMP assays were carried out in triplicate using 500 mL glass bottles with a working volume of 400 mL according to the protocol [47]. Raw and resin-pretreated OMWW was added into the inoculum at a ratio of 0.5 g COD/g vs. (g of OMWW chemical oxygen demand per gram of volatile solids of inoculum). The initial concentration of the inoculum in the BMP test was 9.02 gVS/L. Control samples were prepared without substrate to measure endogenous methane production from the inoculum, which was subtracted from the total methane production of each sample. Another control set used cellulose as substrate to assess inoculum activity. The headspace of each reactor was flushed with nitrogen gas (N₂) for 3 min before sealing with rubber stoppers to ensure anaerobic conditions. The reactors were incubated under mesophilic conditions (35 ± 1 °C) for 35 days in a shaker at a constant rotational speed. The biogas production was determined by pressure measurements using a manometric device (LEO 2, KELLER, Winterthur, Switzerland) according to the ideal gas law, and biogas composition analysis was carried out using gas chromatography (GC) with a CLARUS 480 system (Perkin Elmer, Waltham, MA, USA), as described in Sambusiti et al. [48] The biodegradability calculation is based on the standard theoretical methane potential of 350 mL CH₄/g COD, reflecting the stoichiometric of complete substrate conversion under ideal conditions.

2.6. Modeling Based on the Kinetic Fractionalization of BMPs of Anaerobic Digestion of Raw and Resin Pretreated OMWW

The model developed by Kouas et al. [42] was used to divide the organic matter of the four substrates (j = 1…4) into distinct compartments (i = 1…3) according to their degradation kinetics: rapidly, moderately, and slowly biodegradable sub-fractions. Three assumptions are made:

(i)

At the end of the period considered for the BMP, no biodegradable matter remains. This assumption was supported by the experimental protocol. The tests were conducted over a sufficient duration (35 days), during which methane production was monitored until reaching a plateau. In other words, the degradation of the entire substrate, denoted as S, into various sub-fractions, indicated as Si, based on the corresponding compartments within the organic matter, starts immediately after the addition of the substrate to the reactor, and the slowest fraction is completely degraded at the end of the BMP.

(ii)

The organic matter is categorized into three sub-fractions. This is particularly important for OMWW due to its heterogeneity and phenolic content. Thus, this model structure offers improved mechanistic resolution compared to single-phase kinetic models, first-order or modified Gompertz models. In integrated treatment systems, selective removal of specific fractions (e.g., via adsorption) changes the biodegradability profile. This model allows for a more nuanced assessment of how such pretreatments redistribute organic matter and affect methane production kinetics.

(iii)

The degradation rate of degradation of each sub-fraction (or compartment) remains constant and follows zero-order kinetics (d${\mathrm{S}}_{\mathrm{i}}^{\mathrm{j}}$/dt = −${\mathrm{K}}_{\mathrm{i}}^{\mathrm{j}}$ as long as S_i is positive).

For each fraction “i” of the substrate “j”, this model is parametrized by two parameters: the initial value of the matter in the compartment “i” and the associated kinetics ”${\mathrm{K}}_{\mathrm{i}}^{\mathrm{j}}$”. They were identified from the available data obtained from batch tests using a least-square optimization procedure. For a given substrate “j”, the sum of the three terms ${\mathrm{S}}_{\mathrm{i}}^{\mathrm{j}}$(0), i = 1…3, expressed as a percentage of Sj(0) at t = 0 (the initial quantity of substrate j), indicates the maximum volume of methane that can be generated from the degradation of the total added biodegradable matter. The amount of substrate was quantified in mL of CH₄, which is the volume of methane generated from added organic matter. Therefore, for a given substrate, ${\mathrm{S}}_{\mathrm{i}}^{\mathrm{j}}$ at t = 0, Si(0) is calculated as follows:

$${\mathrm{S}}_{\mathrm{i}}^{\mathrm{j}} (0)=(\mathrm{Quantity} \mathrm{of} \mathrm{COD} \mathrm{added})\times \mathrm{BMP}$$

(5)

S_i at time t (S_i(t)) represents the volume of methane remaining to be produced at time t (mLCH₄) by the sub-fraction Si and was calculated as follows: the degradation kinetics ${\mathrm{K}}_{\mathrm{i}}^{\mathrm{j}}$ were expressed in mL of methane per day and represent the methane production rate from sub-fraction i of substrate J. Zero-order kinetics was considered for the breakdown of organic matter in each compartment. Therefore, the volume of methane produced at time t from each compartment Vol_i (t) can be calculated using Equation (6):

$${\mathrm{V}\mathrm{o}\mathrm{l}}_{\mathrm{i}}^{\mathrm{J}} (\mathrm{t})=\min ({\mathrm{K}}_{\mathrm{i}}^{\mathrm{J}}\times \mathrm{t}, {\mathrm{S}}_{\mathrm{i}}^{\mathrm{J}} (0)) \mathrm{with} \mathrm{i}\in \left[1,2,3\right],\mathrm{J} \mathsf{ϵ}\left[1\dots \mathrm{N}\right]$$

(6)

The minimum operator between ${\mathrm{K}}_{\mathrm{i}}^{\mathrm{J}}$ and ${\mathrm{S}}_{\mathrm{i}}^{\mathrm{J}}$(0) ensures that the volume of methane produced remains within the maximum potential of sub-fraction i.

The total volume of methane produced at time t was as follows:

$$\mathrm{Vol}(\mathrm{t})=\sum {\mathrm{V}\mathrm{o}\mathrm{l}}_{\mathrm{i}}^{\mathrm{J}}(\mathrm{t}) \mathrm{with} \mathrm{i}\in \left[1,2,3\right],\mathrm{J} \mathsf{ϵ}\left[1\dots \mathrm{N}\right]$$

(7)

The residual amount of each sub-fraction at time t can be expressed as follows:

$${\mathrm{S}}_{\mathrm{i}}^{\mathrm{J}}(\mathrm{t})=\max ({\mathrm{S}}_{\mathrm{i}}^{\mathrm{J}} (0)-{\mathrm{K}}_{\mathrm{i}}^{\mathrm{J}}\mathrm{t}, 0) \mathrm{with} \mathrm{i}\in \left[1,2,3\right],\mathrm{J} \mathsf{ϵ}\left[1\dots \mathrm{N}\right]$$

(8)

The max operator ensures that the residual biodegradable fraction remains non-negative, thereby preserving mass balance.

The total quantity of substrate left at t was calculated as follows:

$${\mathrm{S}}^{\mathrm{J}}(\mathrm{t})={\sum }_{\mathrm{j}}{\sum }_{\mathrm{i}}{\mathrm{S}}_{\mathrm{i}}^{\mathrm{J}}(\mathrm{t}) \mathrm{with} \mathrm{i}\in \left[1,2,3\right] \mathrm{and} \mathrm{J} \mathsf{ϵ}\left[1\dots \mathrm{N}\right]$$

(9)

The kinetic model parameters were estimated by non-linear regression using MATLAB R2008. In order to assess the quality of the fit between the model and experimental data, the coefficient of determination (R²) and the Root Mean Square Error (RMSE) were calculated.

3. Results and Discussion

3.1. Effect of Varying Height Bed of XAD-4 Resin Column on Continuous Adsorption Efficiency of OMWW Polyphenolics Compounds

The pH remained consistently at 4.03 ± 0.06 after filtration. The Total Chemical Oxygen Demand (TCOD) decreased from 260 ± 6 g O₂/L to 243 ± 2 g O₂/L, indicating a reduction of 6.35%. Total solids decreased from 145 ± 4 g/L to 138 ± 2 g/L, representing a reduction of 7.72%, while total volatile solids declined from 111 ± 3 g/L to 104 ± 4 g/L, indicating a decrease of 7.1% (

Table 2. Physico-chemical characterization of raw and filtered OMWW (mean (3) ± standard deviation).

Parameters (Unit)	Raw-O	Filtred-O
pH	4.1 ± 0.1	4.03 ± 0.06
TCOD (gO2/L)	260 ± 6	243 ± 2
Total solids (g/L)	145 ± 4	138 ± 2
Total volatile solids (g/L)	111 ± 3	104 ± 4
Total phenolic compounds (g/L)	8 ± 1	7 ± 2

).

The observed decreases show that filtration eliminates particulate matter and some organic substances, resulting in a slight reduction in the overall organic load. Filtration alone was not sufficient to remove dissolved components such as polyphenols, which contribute significantly to the recalcitrance and toxicity of OMWW. This constraint highlights the need for advanced treatment methods, such as adsorption, to effectively manage these compounds. Filtration plays a critical preliminary role in removing solids and facilitating subsequent adsorption processes. This study investigated the effect of varying bed heights on the continuous adsorption efficiency of polyphenolic compounds using an XAD-4 resin column. Three-bed heights (C1: 5.7 cm, C2: 12.1 cm, C3: 18.5 cm) were evaluated for their adsorption capacity to remove polyphenols and organic contaminants (

Table 3. Summary of hydrodynamic and adsorption performance parameters for the three column configurations (C1, C2, C3).

Parameters	C1-5.7 cm	C2-12.1 cm	C3-18.5 cm
Bed Volume Vb (mL)	11.45	24.34	37.18
EBCT (min)	2.29	4.86	7.41
Superficial Velocity (m/s)	4.15·10−4
Reynolds Number (Re)	66
Breakthrough Time (min)	2.25	3.7	8.0
Breakthrough Volume (mL)	4.5	7.4	16
Initial TCOD (g O2/L)	122 ± 2
Final TCOD (g O2/L)	47 ± 4	44 ± 2	24 ± 2
COD removal yield (%)	61	64	80
Polyphenols Initial (mg GAE/L)	3764 ± 88
Polyphenols Final (mg GAE/L)	2275 ± 148	1824 ± 270	1333 ± 388
Total polyphenol adsorption capacity, qtotal (mg)	72.78	233.35	462.60
Adsorption capacity (mg/g)	29	47	47
Polyphenol adsorption yield at saturation (%)	43	56	64
Time to saturation (min)	22.5	55.5	96

).

The present study shows the significant impact of variation in bed height on the adsorption efficiency of polyphenolic compounds from OMWW using XAD-4 resin. The increase in bed height from 5.7 cm to 18.5 cm significantly improved polyphenol removal from 43% to 64%, and TCOD removal from 61% to 80%. The enhancements are mainly due to the increase in resin volume, which provides a greater number of active adsorption sites and extends the contact time between solutes and the adsorbent, which allows a longer diffusion in the porous structure of the resin [47]. The saturation time rose with the increase in bed height, from 22.5 min in C1 to 55.50 min in C2 and reaching 96 min in C3. This trend shows longer residence times and enhanced potential for interactions between the phenolic compounds and the resin. The polyphenols adsorption capacity of the resin (q_e) increased from 29 mg/g in C1 (5.7 cm) to 47 mg/g in C2 and C3 at bed heights of 12.1 and 18.5 cm. Despite column C3 having the longest contact time, its adsorption capacity did not exceed that of C2, which means that the 55.5 min contact time in C2 was sufficient to achieve maximum adsorption efficiency.

The calculated hydrodynamic parameters for the three-column configurations (C1, C2, C3) were used to interpret the column performance (Table 3). The EBCT increased with bed height, reaching 2.29, 4.86, and 7.44 min for C1, C2, and C3, respectively. The superficial velocity remained constant at 4.15 of 10⁻⁴ m/s in all experiments, and the Reynolds number (Re = 66) confirmed laminar flow conditions, which are favorable for equilibrium-based adsorption. Zhang et al. [49] reported that slower saturation allows deeper penetration of the adsorbate into the microporous structure of the resin, which improves overall adsorption performance. Furthermore, Wang et al. [50] documented that beyond a specific limit, larger beds may show a decrease in effective adsorption due to dispersed flow pathways and decreased mass transfer.

The comparison of polyphenol and TCOD removal underscores the non-selective adsorption properties of XAD-4 resin. The treatment aimed to selectively eliminate phenolic compounds, achieving a removal efficiency of 64% for C3. In addition, the total chemical oxygen demand (TCOD) was reduced by 80%, showing that the resin also adsorbed various non-phenolic organic compounds. These results are consistent with the findings of Frascari et al. [45], who observed that under conditions optimized for polyphenol recovery, XAD-type resins adsorbed a significant amount of non-phenolic compounds. In the study of Vavouraki et al. [30], XAD-4 resin removed 65% of total phenolic compounds and reduced absorbed COD from OMWW by about 40%, confirming its non-selective adsorption behavior. The strong adsorption profile is attributed to the macroporous structure and non-specific surface affinity of XAD-4, which facilitates interactions with various constituents in OMWW, including sugars, volatile fatty acids, and other organic materials. This reduction in biodegradable substrates may be detrimental to further methane production in anaerobic digestion. The resin significantly decreased phenolic inhibitors; however, its simultaneous adsorption of biodegradable organics, including sugars and acids, underscores the potential for substrate loss in subsequent anaerobic digestion processes. These results highlight the need to optimize resin selection and dosage in order to achieve a balance between detoxification and energy recovery.

3.2. Effect of Column XAD-4 Resin Adsorption on BMP of Raw and Pretreated OMWW

The implementation of XAD-4 resin columns as a pretreatment for OMWW markedly affected the BMP in terms of NmL CH₄ per g COD over 35 days, the overall biodegradability of the substrate and the total methane volume produced per liter of OMWW, as shown in

Table 4. Effect of pretreatment on methane yield, biodegradability and methane production per liter of OMWW before dilution for raw and XAD-4 resin-pretreated OMWW (mean (3) ± standard deviation).

Samples	Initial TPhCs Concentration (mgTPhCs/L)	Methane Yield (NmL CH4/g COD)	Biodegradability (%)	Biodegradability Enhancement (%)	Methane Production (NLCH4/L of OMWW)
R-O	44.3 ± 0.1	120 ± 1	34	-	31.2
C1	78.1 ± 2.5	267 ± 2	76	+122	25.1
C2	63.2 ± 0.2	287 ± 2	82	+138	25.2
C3	95.4 ± 0.8	272 ± 3	78	+126	6.5

.

The untreated OMWW generated a methane yield of 120 ± 1 NmL CH₄/g COD, corresponding to biodegradability of 34%, highlighting the complexity and low biodegradability of the raw substrate. Pretreatment with adsorption using an XAD-4 resin column led to a significant improvement in the biodegradability of OMWW residual COD. BMP values increased to 267 ± 2 NmL CH₄/g COD for C1 and 272 ± 3 NmL CH₄/g COD for C3, with the maximum yield obtained for C2 at 287 ± 2 NmL CH₄/g COD, corresponding to a 138% improvement over the raw OMWW. The results indicate that adsorption pretreatment effectively improves the biodegradability of residual COD.

The biodegradability improved significantly, reaching 76%, 82%, and 78% for C1, C2, and C3, respectively, compared to only 34% in the untreated substrate. This improvement cannot be attributed to the removal of polyphenol-related inhibition, as all BMP assays were performed at the same COD concentration, resulting in low polyphenol levels that remained significantly below the inhibitory level reported in the literature, generally exceeding 500 mg/L for non-adapted inocula [26]. However, pretreatment by XAD4 resin decreased the total COD, facilitating more efficient anaerobic conversion of the remaining organic matter, especially in the C2 configuration, which presented the most balanced performance.

This shows that adsorbed compounds were more difficult to biodegrade than the remaining ones. Despite the highest bed height and the most extended contact duration, the C3 column did not outperform C2 in methane output, which reached a maximum of 287 NmL CH₄/g COD. Current results demonstrate a higher methane recovery efficiency than reported by Vavouraki et al. [30], who achieved a BMP of 112.8 mL CH₄/g COD for raw OMWW, and only 129.6 mL CH₄/g COD after resin pretreatment. The difference increases when methane production is quantified per liter of OMWW: C2 produced 25.25 NL CH₄/L, but C3 generated only 6.52 L CH₄/L (Table 4).

The significant COD removal observed in C3 shows that the adsorption process not only targeted polyphenolic compounds but also removed a significant amount of readily biodegradable organic matter. This extensive extraction reduced the methane volume per liter of substrate. In contrast, C2 provided a better balance between polyphenol removal and preservation of biodegradable compounds, yielding the highest methane yield. However, the total methane produced per liter of OMWW remained lower than that of the raw OMWW due to the reduced organic load after adsorption. Bovina et al. [31] reported maximum methane yields of 750 mL CH₄/g VS for dephenolized OMWW and 470 mL CH₄/g VS for raw OMWW. When expressed per liter of treated wastewater, the yields were 9.98 L CH₄/L and 11.99 L CH₄/L, respectively, which are significantly lower than the 25.25 L CH₄/L attained under the C2 condition in this study. This contrast highlights the importance of pretreatment strategies that both reduce inhibitory compounds and maintain biodegradable organic matter. Bovina’s study demonstrated that the adsorption using XAD-16 resin pretreatment resulted in a 32.2% decrease in dissolved COD and a 47.8% in volatile solids, suggesting that a considerable amount of digestible substrate was removed in addition to polyphenols. This approach improved stability but ultimately decreased volumetric methane recovery. In comparison, the C2 configuration in this study demonstrated a more advantageous equilibrium between detoxification and substrate conservation, leading to higher methane potential and improved energy recovery per unit volume of OMWW.

Improved efficiency of anaerobic digestion and the recovery of polyphenols highlight the potential of adsorption as a sustainable approach for the valorization of OMWW. In addition to supporting bioenergy recovery, the process enables the extraction of polyphenolic compounds with significant economic and environmental value, particularly for applications in the pharmaceutical, cosmetic, and food industries. These benefits require further study under continuous-flow conditions, as increased polyphenol concentrations may influence anaerobic digestion performance, selectivity, and stability. Such validation is crucial to establish the robustness and scalability of the proposed approach in the context of a circular economy.

3.3. Compartmentalization of the Organic Matter Contained in Raw and Pretreated OMWW Based on the Kinetics of Anaerobic Digestion

A kinetic modeling approach, adapted from the model developed by Kouas et al. [42], is used in this section to examine how pretreatment with XAD-4 resin alters the distribution and degradation behavior of organic matter in OMWW. The model divides total organic matter into three fractions: rapidly biodegradable (S1), moderately biodegradable (S2), and slowly biodegradable (S3), each characterized by its own specific methane production rates (K1, K2, K3). The quality of the model adjustment was assessed using standard statistical indicators. The R² was 0.99 for the Raw-O, C2, and C3 configurations, indicating an excellent fit, while C1 exhibited a lower R² of 0.68, suggesting some deviation from the model. The RMSE values further supported these observations, with a low error of 4.87 for Raw-O, 7.60 for C2, and 7.69 for C3, whereas C1 displayed a notably higher RMSE of 61.18.

Table 5. Kinetic parameters for the compartmentalization of organic matter from the monodigestion of raw and XAD-4 resin-treated OMWW.

Substrate	Parameters
	S1(0) %	K1 (mL CH4/d)	S2(0) %	K2 (mL CH4/d)	S3(0) %	K3 (mL CH4/d)
Raw-O	74	1.82	13	0.07	13	0.02
C1	39	5.49	43	20.68	18	1.38
C2	34	18.08	34	9.41	33	2.68
C3	50	23.86	13	2.47	38	2.92

presents a summary of the kinetic parameters derived from raw OMWW and the three pretreated conditions (C1, C2, and C3).

Figure 1 and Figure 2 present the results of kinetic modeling and the progression of biodegradable sub-fractions during the anaerobic digestion process. Kinetic modeling of raw OMWW showed that, despite a high content of rapidly degradable substrate (S1 = 74%), the methane production rate was relatively low (K1 = 1.82 mL CH₄/d). This suggests that the limited microbial access or the complexity of the matrix may have hindered degradation. Similarly, the moderately (S2 = 13%) and slowly (S3 = 13%) degradable fractions had low methane generation rates (K2 = 0.07 and K3 = 0.02 mL CH₄/d), which confirms that a significant portion of the organic matter remained poorly accessible (Figure 1).

The results highlight how the distribution of organic matter influences methane production and degradation kinetics. The use of XAD-4 resin improved substrate accessibility, enhanced methane production rates, and restructured the distribution of biodegradable fractions. Figure 2 shows model calibration and sub-fraction evolution for C1, C2, and C3, demonstrating consistent improvements in methane production as resin bed height increased.

In C1, the methane production kinetics increased compared to raw OMWW (K1 = 5.49 mL CH₄/d) despite lower S1 content (39%). The presence of a larger moderately degradable fraction (S2 = 43%) with a high degradation rate (K2 = 20.68 mL CH₄/d) reflects better conversion of solubilized organic matter. In C2, methane production kinetics improved significantly (K1 = 18.08 mL CH₄/d) despite a lower S1 fraction (34%), indicating enhanced biodegradability. Organic matter was evenly distributed across the three fractions (S1 = 34%, S2 = 34%, S3 = 33%) and high kinetics constants (K2 = 9.41, K3 = 2.68 mL CH₄/d). This balanced profile indicates that the C2 pretreatment condition retains accessible, moderately complex, and slowly degradable substrates, resulting in effective and sustained methane production.

C3 had the highest methane production rate (K1 = 23.86 mL CH₄/d) and a high rapidly degradable fraction (S1 = 50%). However, the moderate fraction (S2 = 13%) was reduced, with a lower degradation rate (K2 = 2.47 mL CH₄/d), suggesting that excessive adsorption may have eliminated not only recalcitrant compounds but also valuable intermediates. The high rate of S3 degradation (K3 = 2.92 mL CH₄/d) and the increased S3 fraction (38%) indicate better accessibility to more recalcitrant components. Although 80% of COD was removed during pretreatment, the remaining biodegradable matter showed significantly higher degradation rates, resulting in improved methane conversion. The analysis confirms that resin pretreatment modifies the structure of the substrate and kinetics. In particular, C2 achieved a methane yield of 287 ± 2 NmL CH₄/g COD, with enhanced rates across all degradation fractions, confirming the operational advantage of selective removal of compounds.

These results indicate that resin-based pretreatment restructures the organic matter matrix and improves kinetic performance. The improvements observed in this study were not linked to the polyphenol detoxification, as concentrations remained below inhibitory limits. The enhancement of anaerobic digestion performance may result from the selective removal of poorly biodegradable compounds, which improves the accessibility of the remaining organic matter for microbial degradation. These results highlight the double impact of pretreatment: on the one hand, it enhances methane production by improving microbial accessibility and reducing toxicity; on the other hand, a too aggressive removal of organics can compromise substrate quality and reduce overall methane production. This highlights the need for precise control and optimization of the selection of the resin and of bed height and flow parameters.

4. Conclusions

This study examined a novel integrated approach for valorization OMWW by combining Amberlite XAD-4 resin adsorption with anaerobic digestion. The process significantly improved treatment efficiency, eliminating up to 64% polyphenol, 80% COD, and increasing biodegradability from 34% to 82%. Among the tested conditions, the C2 resin bed height increased the methane yield by approximately 138% (287 ± NmL CH₄/g COD) compared to untreated OMWW. The adsorption pretreatment enhanced substrate accessibility by removing poorly biodegradable fractions while allowing partial recovery of valuable phenolic compounds. However, higher resin loads resulted in a decrease in methane yield per liter of OMWW, indicating a trade-off between detoxification and biogas production. The C2 configuration achieved the best balance between these outcomes.

This performance highlights the benefit of selectively removing poorly biodegradable and inhibitory compounds while preserving bioavailable organics for anaerobic digestion. These findings support the implementation of integrated treatment systems aligned with circular economy objectives in Mediterranean olive mills. However, this study was limited to batch-scale anaerobic digestion operation, and the long-term performance under continuous-flow conditions remains to be validated. Additionally, the cost-effectiveness of the resin and the potential for fouling or performance decline over extended use require further investigation. Future research should explore more selective adsorbents suitable for OMWW composition, especially for matrices with higher COD concentrations, and evaluate the integration of this process into continuous-flow pilot-scale systems to enhance scalability and operational stability.