Comparative Life Cycle Assessment of Physical and Chemical Activation Routes for Oil Palm Shell-Derived Activated Carbon in Lufenuron 50-EC Pesticide Adsorption

Abstract

This study evaluates the life cycle assessment (LCA) of Lufenuron 50-EC pesticide adsorption from aqueous solution using oil palm shell (OPS)-derived activated carbon produced through two activation routes: physical and chemical. The assessment covers environmental impacts associated with feedstock collection, transportation, pre-processing, and post-processing stages involved in producing activated carbon for pesticide removal. The cradle-to-grave LCA technique was applied using the ELCD 3.2 Greendelta v2.18 database and processed with OpenLCA v2.4 using CML-IA baseline method to perform the quantitative life cycle impact assessment. The results for treating 1 m³ of contaminated water show that physical activation route (Route 1) generates a higher environmental burden across all evaluated impact categories compared to chemical route (Route 2). Notably, global warming potential (GWP) reached 117.62 kg CO₂ eq for Route 1 compared to 75.86 kg CO₂ eq for Route 2. This represents a 35.5% reduction with the chemical route, suggesting that the high energy demand associated with thermal process in physical activation generates more significant greenhouse gas emissions. Overall, this study helped identify critical performance points and opportunities for improvement in converting the OPS to an activated carbon transformation process and its application in pesticide contamination control.

1. Introduction

Pesticides are agrochemicals widely used in agriculture to prevent and control pests (insects, fungi, rodents, and unwanted plant species) that can affect farmers’ crop [1]. The World Health Organization (WHO) classifies pesticides according to their hazard into five categories, ranging from category IV/5, whose warning phase is precautionary, to Ia/1 or extremely hazardous, with a very toxic warning phase. The latter are the most dangerous for life in general [2]. Although the use of agrochemicals is often perceived as a necessity, their excessive implementation has become a critical environmental problem. This problem arises from the lack of viable and ecologically friendly alternatives to current practices, which generate serious impacts on multiple environmental matrices. These impacts include the dispersion of particles to neighboring areas, soil fertility, reduction in the biodegradation capacity, destabilization of ecological integrity, and water quality, among others, reducing the survival capacity of native species in the affected ecosystems [3,4,5,6,7,8]. Likewise, this practice causes serious illnesses on human health, some of which can even cause the death of the affected person. These illnesses include: cancer, respiratory diseases, Alzheimer’s, diabetes, and Parkinson’s, among others [9]. Lufenuron 50-EC is a pesticide widely used in oil palm crops. It belongs to the group of insecticides that acts as an inhibitor of chitin synthesis, regulating insect growth [10]. Previous studies [11,12] have shown how the accumulation of this compound in water bodies causes the population decline of endemic species, irreversibly altering local ecological balances. These findings are particularly alarming when considering the presence of this pesticide in the environment and its ability to accumulate in food chains. Oil palm shell (OPS) is a byproduct of the oil palm processing technique that is generally considered a waste product with no value beyond its use as boiler fuel. However, previous studies have shown great potential for its transformation into activated carbon and subsequent application in the removal of dissolved contaminants in different matrices, such as heavy metals and pesticides [13,14]. Activated carbon is a material recognized worldwide for its exceptional capacity to adsorb contaminants dissolved in liquid and gaseous media due to its large surface area and abundant porosity [15,16,17]. Adsorption is a widely known and used methodology due to its economic and simplicity [18]. It is considered as a surface phenomenon that consists of the retention of particles (adsorbate) on the solid surface (adsorbent), and the amount of adsorbate retained will depend on the physical and/or chemical interactions between the adsorbate and the adsorbent [19,20,21]. Lufenuron 50-EC adsorption using activated carbon produced from oil palm shell, as explored in our previous work [22], achieves efficiencies of up to 97% with activated carbon produced by physical and chemical activation with KOH as the impregnating agent. However, its environmental sustainability remains unexplored, particularly in terms of CO₂ eq emissions associated with KOH activation, a process known for its energy intensity [23].

Life Cycle Assessment is a standardized methodology ISO 14040-44 [24] that assesses the environmental impacts associated with a product, process, or service, considering all stages of its life cycle: from the extraction of raw materials to its final disposal. This methodology quantifies energy flows, emissions, and resource consumption using computational models supported by specialized software (e.g., OpenLCA, SimaPro) and robust databases (e.g., ecoinvent, ELCD, Agribalyse) [25,26], allowing for the identification of critical environmental issues and the comparison of sustainable alternatives.

In the context of applying LCA to adsorption processes with agricultural waste, several studies have been developed using OpenLCA v2.4 software as a tool for modeling complex agricultural waste valorization processes, confirming the advantages of these biosorbents subject to the optimization of processes with high energy demand. For example, Pereira et al. (2024) [27] performed a cradle-to-gate LCA of activated carbon produced from banana peel by chemical activation (NaOH) for the removal of methylene blue dye, reporting a favorable environmental profile compared to other conventional production systems, highlighting significant reductions in impact categories such as dependence on non-renewable resources, land use, and global warming potential. Similarly, Maiti and Meikap (2025) [28] evaluated the environmental impact of the activated carbon production process from green coconut shells for the adsorption removal of Pb (II). Their findings revealed favorable results across multiple impact categories, validating it as a sustainable and environmentally friendly process. Furthermore, Ren et al. (2025) [29] conducted a life cycle assessment of the activated carbon production process derived from poplar branches treated with deep eutectic solvent (DES) for effective biomass fractionation and its subsequent application in the adsorption of methylene blue dye. Their findings revealed that DES treatment could potentially reduce the impact of the subsequent activation process, specifically in environmental impact categories such as energy consumption and environmental degradation.

However, although recent studies have applied LCA to the production of activated carbon derived from different precursors (e.g., vine shoots, banana peel, or waste plastic) for the adsorption of contaminants such as heavy metals, dyes, or hazardous antibiotics [27,30,31], there is a critical gap in the literature: the environmental assessment of specific systems designed for the removal of Lufenuron 50-EC, a widely used pesticide in agriculture with potential ecotoxicological risk. This omission limits the ability to make informed decisions about the management of its environmental impacts, especially in regions where its use is intensive.

This study represents an environmental assessment of pesticide management technologies, as it presents the first life cycle assessment for Lufenuron 50-EC removal using OPS-derived activated carbon. A realistic system is analyzed based on technical parameters previously validated under laboratory conditions [22], including experimental data on adsorption efficiency. The result of this LCA will allow for a detailed quantification of the environmental impacts at all stages of the process, from raw material collection to final disposal (cradle to grave). Furthermore, this analysis will establish a framework with two main applications: (first) guiding the optimization of pilot-scale treatment systems for Lufenuron 50-EC-contaminated water using OPS-derived activated carbon, and (second) assessing the environmental feasibility of its implementation at an industrial scale by analyzing the trade-offs between its high adsorption capacity and the associated carbon footprint. This environmental profile will provide a crucial complementary perspective to the existing technical comparison of various adsorbents derived from agricultural wastes (e.g., peanut shells, olive pomace, macadamia shells) for pesticide removal, enabling a comprehensive sustainability assessment.

2. Materials and Methods

The activation (physical and chemical) and adsorption processes were developed in a previous work [22].

Table 1. Adsorption yield of Lufenuron 50-EC using activated carbon produced from oil palm shell.

Activation Process	Sample	Activation Temperature (°C)	Residence Time (h)	Impregnation Ratio (KOH/Biochar)	Adsorption Yield (% w/w)
Physical	AC-800-2	800	2.0	N/A &	90.15
	AC-900-2	900	2.0	N/A	96.93
Chemical	AC-750-1.5-3:1	750	1.5	3:1	94.01
	AC-800-1-2:1	800	1.0	2:1	94.87

^& Does not apply.

presents the operating parameters for the production of the four ACs selected for characterization and evaluation, as well as their respective adsorption yields (%) of Lufenuron 50-EC in aqueous solution. In the particular case of chemical activation, KOH was used as the impregnation chemical reagent.

The complete characterization of the activated carbons (BET, FTIR, XRD, Raman, SEM) was reported in our previous work [22]. Briefly, four activated carbons were selected (two per activation method) based on their adsorption performance (Table 1). For the adsorption experiments, five solutions were prepared for each activated carbon, varying the adsorbent concentration (0.05–0.45 g) at constant Lufenuron 50-EC concentration (10 ppm) using a particle size of 0.85–2 mm. The mixtures were stirred at 170 rpm for 20 h to reach equilibrium [22].

LCA is a methodology recognized by [24] that systematically assesses the environmental impacts of a product or process at all stages. Its structured approach comprises four key phases or stages that ensure scientific rigor and practical applicability for sustainable decision-making: (1) definition of objectives and scope, (2) inventory of material and energy flows, (3) quantitative assessment of environmental impacts, and (4) critical interpretation of results.

The objective (first stage) of the LCA in this study was to assess the environmental impacts associated with the adsorption of the pesticide lufenuron 50-EC using OPS-derived activated carbon. The scope of this study was defined to include the processes of OPS extraction, transportation, pretreatment, activation (physical and chemical), lufenuron 50-EC adsorption, and final disposal. The system boundaries were selected following the cradle-to-grave approach (Figure 1).

Figure 1 presents a schematic representation of the system boundaries for the two activation routes. The dashed box encloses all unit processes included in the LCA model. Each stage shown in Figure 1 corresponds to specific input and output flows reported in the life cycle inventory: OPS collection (includes electricity and water consumption for collecting oil palm shells from the processing facility), transportation [accounts for the diesel-powered transport of OPS from the collection point to the processing facility (distance expressed in t·km)], grinding and sieving (includes electricity consumption for particle size reduction and selection of the 0.85–2 mm fraction), devolatilization (Route 2 only, includes electricity consumption and N₂ supply for thermal pretreatment of OPS prior to chemical activation), impregnation with KOH (Route 2 only, includes KOH, distilled water, and electricity inputs for impregnation of biochar), activation (includes electricity, CO₂, and N₂ for physical activation, and electricity and CO₂ for chemical activation), washing with distilled water (Route 2 only, includes distilled water for removal of residual KOH from activated carbon), drying in oven (Route 2 only, includes electricity for oven drying of washed activated carbon), lufenuron 50-EC adsorption (includes activated carbon, electricity, and contaminated water as inputs, with treated water and depleted activated carbon outputs), and final disposal (includes depleted activated carbon as an output with no energy recovery or recycling credits assumed).

The dashed box in Figure 1 indicates the system boundary, which follows a cradle-to-grave approach from OPS collection to final disposal of depleted activated carbon. Processes outside the system boundary (e.g., production of background materials such as KOH, CO₂, and N₂) are included via the ELCD 3.2 database but are not shown in the figure for simplicity.

To ensure consistency, scalability, and relevance in the environmental performance assessment, the functional unit (FU) was defined as the treatment of 1 m³ (1000 kg) of water contaminated with Lufenuron 50-EC at an initial concentration of 10 mg/L, achieving a final concentration below 0.5 mg/L (i.e., >95% removal efficiency, as experimentally verified in our previous work [22]). This FU explicitly links the environmental burdens to the service provided: the removal of approximately 9.5 g of Lufenuron 50-EC per m³ of treated water, based on the average adsorption yields reported in Table 1 (93.54% for physical activation and 94.44% for chemical activation).

The life cycle inventory is reported per activation Route (physical and chemical) based on the laboratory-scale treatment of 0.5 kg of contaminated water, corresponding to the experimental conditions. To obtain the impact assessment results per functional unit, the laboratory-scale impacts were linearly scaled using a factor of 2000 (1000 kg/0.5 kg). The linear scaling approach (factor of 2000) is a simplifying assumption. However, it is important to acknowledge that direct linear extrapolation of life cycle inventory data from laboratory to industrial scale is generally considered inappropriate in rigorous prospective LCA studies, as laboratory operations involve discrete, unconnected steps and equipment that differ substantially from commercial-scale facilities [32,33]. Although experimental conditions of this study differ between routes, the same scaling factor was applied to both activation routes, introducing a common systematic bias. Given the comparative objective of this study, this common bias preserves the relative ranking of environmental impacts. Nevertheless, absolute impact values should be interpreted as laboratory-scale projections, not industrial-scale predictions. Future work should employ non-linear scale-up models. This functional unit ensures the standardization of the life cycle inventory and the results associated with the impacts, allowing for the interpretation and direct comparison of data with alternative water treatment technologies, acknowledging that the absolute values presented herein are laboratory-scale projections.

To compile the life cycle inventory (second stage), all input and output flows of the system were determined using primary data from the experimental study [22] and, where applicable, background data from the ELCD 3.2 Greendelta v2.18 database (European reference Life Cycle Database). This database was selected because it is freely available, fully compatible with OpenLCA v2.4, and provides robust characterization factors for the CML-IA method. The geographical scope of the background datasets is the European average, as ELCD primarily represents European processes. No other databases (e.g., ecoinvent, USLCI) were used. The electricity mix used in the inventory corresponds to the Colombian national grid, as the experiments were conducted in Colombia. Since ELCD does not include a specific Colombian electricity dataset, the European average electricity mix from ELCD was used as a proxy. Regarding cut-off rules, an attributional modeling approach was applied. For the end-of-life (EOL) stage, a simplified assumption was adopted: the depleted activated carbon (containing adsorbed Lufenuron 50-EC) was treated as a stabilized output flow without allocation of specific landfill burdens. This simplification is justified by the high stability of Lufenuron 50-EC once adsorbed onto the activated carbon surface, as demonstrated by the strong chemical interactions (carboxylic and hydroxyl groups) identified in the FTIR analysis of our previous work [22], which indicates negligible leaching potential. No further environmental emissions from degradation or desorption were assumed. All other stages (feedstock collection, transportation, pre-processing, activation, and adsorption) were fully modeled without cut-offs. No recycling or end-of-life credits were considered.

All background flows (electricity, transport, water, KOH, CO₂, N₂, waste treatment, etc.) were sourced exclusively from the ELCD 3.2 database. Foreground data (OPS collection, grinding, activation parameters, adsorption yields, etc.) were taken directly from the experimental study [22] in which 18 activated carbons were synthesized from oil palm shell using physical and chemical activation Routes. Four activated carbons were selected for the LCA: the two with the highest Lufenuron 50-EC adsorption capacity from the physical activation Route and the two from the chemical activation Route. Life cycle inventory, environmental impact and contribution analyses are presented by activation Route (physical vs. chemical), aggregating the burdens of the two selected carbons per Route. Therefore, the results represent the best possible scenario for each Route based on the experimental data from [22].

During the life cycle impact assessment (third stage) strictly linked to the system inputs and outputs, an attributional modeling approach [31] was implemented, which allowed for proportionally assigning environmental loads to each experimental stage and facilitated detailed monitoring of the activation (physical and chemical) and adsorption processes. The specialized free access software OpenLCA v2.4 in compliance with the CML-IA baseline method was used to perform the quantitative assessment of life cycle impacts. Midpoint allows for the quantification of environmental impacts, reducing ambiguity and increasing transparency compared to Endpoint. This work considered ten categories (global warming power, abiotic depletion, acidification, eutrophication, fresh water aquatic ecotoxicity, human toxicity, marine aquatic ecotoxicity, ozone layer depletion, photochemical oxidation, and terrestrial ecotoxicity) of common impacts that allowed for the quantification of the environmental burdens associated with each raw material and process in the adsorption of lufenuron 50-EC using OPS-derived activated carbon.

The interpretation (fourth stage) of the results ensured the identification of critical points from activated carbon production to final disposal, allowing for not only prioritizing optimization strategies but also facilitating the establishment of quantitative bases for designing more sustainable experimental protocols.

A univariate sensitivity analysis was performed following ISO 14044 [24] to assess the robustness of the LCA results. Three parameters were selected based on their contribution to total impact and inherent variability: (I) electricity consumption during activation (±20%), (II) Lufenuron 50-EC adsorption yield (±5 percentage points), and (III) transport distance of OPS (±50%). The Global Warming Potential (GWP) was used as the reference impact category. Each parameter was varied independently while holding others at baseline values. The percentage change in GWP relative to baseline was calculated using Equation (1). For the adsorption yield sensitivity, GWP was additionally expressed per gram of Lufenuron 50-EC removed to isolate the effect of yield.

$$Percentagechange={\displaystyle \frac{{GWP}_{scenario}-{GWP}_{baseline}}{{GWP}_{baseline}}}\ast 100$$

(1)

where:

GWP_scenario = GWP value for the low or high scenario (kg CO₂ eq per functional unit).

GWP_baseline = GWP value for the baseline scenario (kg CO₂ eq per functional unit).

3. Results

3.1. Activated Carbon Characterization

The complete characterization of the activated carbons (including BET surface area, FTIR, XRD, Raman, and SEM analyses) was reported in our previous work [22]. Key findings relevant to this LCA study are that the activated carbons exhibited adequate surface area (421–548 m²/g for physical activation; 22–90 m²/g for chemical activation), favorable surface chemistry (including hydroxyl and carboxyl groups), and a porous morphology suitable for Lufenuron 50-EC adsorption. Chemical activation with KOH introduced additional carboxylic groups and resulted in more pronounced pore development compared to physical activation. For full details, the reader is referred to [22].

Table 2. Adsorption capacity of pesticides using activated carbons from agro-industrial waste.

Raw Material	Adsorbent	Pesticide	Adsorption Capacity (mg/g)	Reference
Silver berry seeds	Activated carbon	Acetamiprid	193.92	[34]
Peanut shell	Activated carbon	Methomyl	56.62	[35]
Coffee waste	Activated carbon	Lindane	3.80	[36]
Olive oil pomace	Activated carbon	Methomyl	277.30	[37]
		Imidacloprid	233.97
		Metalaxyl	119.71
		Paraquat Dichloride	74.94
Peach stones	Activated carbon	2,4-D	496.81	[38]
Prickly pear seeds	Activated carbon	Chlorpyrifos	35	[39]
Industrial sludge	Activated carbon	Imidacloprid	153.10	[40]
		Thiamethoxam	124.40
		Acetamiprid	124.30
Macadamia nut shells	Activated carbon	2,4-D	600	[41]
Walnut shells	Activated carbon	2,4,5-T	224.60	[42]
		2,4-D	290.20
Oil palm shell	AC-800-2 (Physical)	Lufenuron 50-EC	1011	[22] Results for the current study
Oil palm shell	AC-900-2 (Physical)	Lufenuron 50-EC	1352
Oil palm shell	AC-750-1.5-3:1 (Chemical)	Lufenuron 50-EC	1167
Oil palm shell	AC-800-1-2:1 (Chemical)	Lufenuron 50-EC	1149

presents a comparative summary of the adsorption capacities (mg/g) recently reported in the literature for activated carbon produced from different raw materials for the removal of various pesticides.

Numerous studies have been conducted on the production of activated carbon from lignocellulosic biomass for use in pesticide removal in aqueous matrices [43,44,45]. Among the raw materials listed in Table 2, oil palm shell (OPS) shows the highest adsorption capacity for the removal of Lufenuron 50-EC.

The adsorption values reported for OPS are in the range of 1011 to 1352 mg/g, corresponding to the author’s previous study [22] and constituting the technical and experimental support of the present work. These adsorption capacities are significantly higher than those reported for other raw materials, including silver berry seeds (193.92 mg/g), coffee waste (277.90 mg/g), peach stones (496.81 mg/g), macadamia nut shells (600 mg/g), and walnut shells (290.20 mg/g). This stark contrast highlights the superior affinity between OPS-derived activated carbon and the pesticide Lufenuron 50-EC, emphasizing the effectiveness of the production processes employed.

The market adsorption performance of these OPS-derived activated carbons is attributed to their high internal surface area and porous structure dominated by mesopores. Furthermore, they are enhanced by their favorable chemical interaction with the adsorbent, primarily due to surface hydroxyl and carbonyl groups. In the specific case of chemically activated carbons, an additional contribution arises from carboxylic groups introduced during KOH impregnation. Kinetic analysis confirmed that the pseudo-second-order model best represents the experimental data (R² = 0.91–0.99), indicating that chemisorption governs the adsorption process [46].

3.2. Life Cycle Inventory

Table 3. Life cycle inventory of Lufenuron 50-EC adsorption using OPS-derived activated carbon per functional unit (1 m³ of treated water).

Stages	Input	Physical Activation	Chemical Activation
		Amount	Amount
OPS collection	Electricity	1660 kWh	1660 kWh
	Water	110,000 kg	110,000 kg
Transportation	OPS	66,000 kg	66,000 kg
	Distance	79,200 t·km	79,200 t·km
Grinding and sieving	OPS	6600 kg	6600 kg
	Electricity	2480 kWh	2480 kWh
Devolatilization	OPS	N/A &	320 kg
	Electricity	N/A	34,240 kWh
	N2	N/A	560 kg
Impregnation with KOH	Biochar	N/A	80 kg
	KOH	N/A	40 kg
	Electricity	N/A	2640 kWh
	Distilled water	N/A	40 kg
Activation	OPS	320 kg	N/A
	KOH-impregnated biochar	N/A	140 kg
	Electricity	79,240 kWh	50,000 kWh
	CO2	220 kg	240 kg
	N2	571 kg	N/A
Washing with distilled water	KOH-impregnated activated carbon	N/A	116.7 kg
	Distilled water	N/A	4000 kg
Dried in oven	Wet activated carbon	N/A	86.6 kg
	Electricity	N/A	144,000 kWh
Lufenuron 50-EC adsorption	Activated carbon	5 kg	5 kg
	Electricity	9600 kWh	9600 kWh
	Contaminated water	1000 kg	1000 kg
Final disposal	Depleted activated carbon	5 kg	5 kg

^& Does not apply. The stages shown in this table correspond to the system boundary stages illustrated in Figure 1.

shows the flows associated with the life cycle inventory of the Lufenuron 50-EC adsorption process using activated carbon produced from OPS by physical and chemical activation.

It should be noted that for the final disposal stage, no specific background process (e.g., landfilling or incineration) from the ELCD 3.2 database was assigned to the depleted activated carbon flow. Instead, the depleted carbon was treated as a stabilized output, consistent with the simplified EOL assumption described above. This approach is conservative regarding leaching potential, as it assumes no environmental release from the stabilized material, but it may underestimate impacts associated with long-term storage or final disposal infrastructure.

The purpose of the life cycle inventory was to assess the environmental impacts associated with the adsorption process of the pesticide Lufenuron 50-EC dissolved in synthetic solutions, using OPS-derived activated carbon as the adsorbent. Specifically, two activation processes or methods (physical and chemical) were evaluated.

Unlike previous studies [23,26] in which the LCA system is limited to the activated carbon production process, excluding the adsorption and final disposal stages despite the fact that activated carbon is used in adsorption processes, the LCA system in this study is limited by all process stages ranging from raw material collection to final disposal of depleted activated carbon (Figure 1), including both activation routes, the adsorption process and final disposal. Table 3 reveals several key differences between the two activation routes. While the upstream stages (OPS collection, transportation, grinding/sieving) are identical for both routes, the activation stage differs substantially. Physical activation consumes considerably more electricity (79,240 kWh) than chemical activation (50,000 kWh) but requires no chemical inputs beyond CO₂ and N₂. In contrast, chemical activation involves a multi-step process: devolatilization (34,240 kWh), KOH impregnation (2640 kWh), activation (50,000 kWh), washing with 4000 kg of distilled water, and oven drying (144,000 kWh). Notably, the drying stage consumes the highest amount of electricity (144,000 kWh) but, as shown later in the contribution analysis (Section 3.4), its relative environmental impact is negligible due to the dominance of the adsorption stage. The adsorption stage itself consumes 9600 kWh for both routes, using 5 kg of activated carbon to treat 1 m³ of contaminated water. This high energy demand is further discussed in Section 3.4 as a key limitation of the laboratory-scale batch configuration.

3.3. Environmental Impact Assessment

Table 4. Environmental impacts per 1 m³ of water treated (10 mg/L of Lufenuron 50-EC).

No.	Impact Category	Abb.	Unit	Activation Route
				1 (Physical)	2 (Chemical)
1	Global warming potential	GWP	kg CO2 eq	117.62	75.86
2	Abiotic depletion (fossil fuels)	AD	MJ	527.50	339.82
3	Acidification	AC	kg SO2 eq	0.18	0.11
4	Eutrophication	Eut	kg PO4 eq	0.05	0.02
5	Fresh water aquatic ecotox.	FWAE	kg 1,4-BD eq	0.19	0.13
6	Human toxicity	HT	kg 1,4-BD eq	−5.62	−4.42
7	Marine aquatic ecotoxicity	MAE	kg 1,4-BD eq	4592.10	2806.86
8	Ozone depletion potential	ODP	kg CFC-11 eq	9.13 × 10−7	3.42 × 10−7
9	Photochemical oxidation	PO	kg C2H4 eq	0.02	0.01
10	Terrestrial ecotoxicity	TE	kg 1,4-BD eq	0.01	3.33 × 10−3

presents the results of the environmental impacts associated with the treatment of 1 m³ of synthetic water with a concentration of 10 mg/L of adsorbate (Lufenuron 50-EC), using activated carbon produced by physical and chemical activation (Route 1 and 2, respectively).

The original impact results were calculated for 0.5 kg of contaminated water per activation Route (physical and chemical), corresponding to the laboratory-scale experiments. To improve comparability and follow ISO 14040/44 [24] recommendations, these results were scaled to the functional unit of 1 m³ (1000 kg) of treated water. The scaling factor was calculated using Equation (2), and each impact category result was multiplied by the scaling factor (2000), obtaining the values in Table 4.

$$Scalefactor={\displaystyle \frac{1000\mathrm{k}\mathrm{g}\left(funtionalunit\right)}{0.5\mathrm{k}\mathrm{g}\left(laboratoryscaleperRoute\right)}}=2000$$

(2)

In terms of GWP, physical activation (Route 1) recorded an impact of 117.62 kg CO₂ eq compared to 75.86 kg CO₂ eq for chemical activation (Route 2). This represents a 35.5% reduction in global warming potential when using the chemical Route instead of the physical Route, suggesting that the high energy consumption of the physical route’s thermal process generates more significant greenhouse gas emissions. A previous study [47] implemented a functional unit of 1 g of adsorbed pollutant in an LCA related to water treatment using adsorbents (activated carbon and biochar) from agricultural waste, obtaining GWP values between 5.4 and 9.8 kg CO₂ eq. A direct comparison of absolute GWP values with the literature is challenging due to differences in functional units. Nandikes et al. [47] reported GWP values ranging from 5.4 to 9.8 kg CO₂ eq per gram of adsorbed pollutant for various agricultural waste-derived adsorbents. Although a direct quantitative comparison with the results of this study is not appropriate due to differences in system boundaries, adsorbate, and scale, it is noteworthy that the relative ranking of the two routes evaluated in this study (chemical activation outperforming physical activation) is consistent with the trends reported in the literature, where chemically activated carbons often exhibit lower environmental impacts per gram of pollutant removed [27,47].

Another similar study [48], in which 1 m³ of produced drinking water was chosen as the functional unit in an LCA for drinking water production using granular activated carbon as a filter for contaminant removal, obtained GWP values (0.515 kg CO₂ eq) well below those obtained in this study (117.62 and 75.86 kg CO₂ eq for Route 1 and 2, respectively). This substantial disparity is primarily attributed to the laboratory-scale batch configuration used in this study. As shown in Table 3, the adsorption stage consumes 9600 kWh per functional unit (1 m³ of treated water). This value was obtained by linearly scaling the laboratory experimental data (which included mechanical stirring at 170 rpm for 20 h) to the functional unit. In contrast, Klimtová et al. [48] modeled an industrial-scale continuous column system, where energy consumption is dominated by pumping (typically 0.5–5 kWh per m³). The linear scaling applied in this study amplifies the laboratory energy demand, whereas real industrial processes benefit from economies of scale and energy efficiencies not captured by linear extrapolation. Other factors, such as differences in the aqueous matrix (raw river water vs. synthetic pesticide solution) and adsorbate characteristics, may also contribute but are secondary to the effect of scale and system configuration.

In the AD impact category, the physical route also reveals greater consumption of fossil resources (527.50 MJ) compared to the chemical route (339.82 MJ), implying a difference of 55.2%, which reaffirms the high energy demand of the physical process. For the AC impact category, the physical route generated 0.18 kg SO₂ eq, compared to 0.11 kg SO₂ eq by the chemical route; this suggests a significant reduction (38.9% less) in acidifying emissions with the chemical route. The impact due to eutrophication was more relevant in the physical route (0.05 kg PO₄ eq) compared to the chemical route (0.02 kg PO₄ eq). This may be related to the consumption of CO₂ as an activating gas; although CO₂ is not directly eutrophicating, its industrial production, compression or transportation may be associated with emissions of nitrogen oxide (NO_x), which contributes to eutrophication once it is deposited in water bodies. Furthermore, energy dependence also plays an important role since the physical activation process involves a higher electrical energy consumption in the activation stage (39.62 kWh) compared to chemical activation (25 kWh), which can be associated with higher indirect NO_x emissions considering that the energy matrix in Colombia is dependent on fossil fuels. In the case of the FWAE impact category, the physical route (0.19 kg 1,4-BD eq) presented a slightly greater impact than the chemical route (0.13 kg 1,4-BD eq), which represents a 31.5% reduction in the impact on fresh water ecosystems for the chemical route.

For the HT impact category, both activation routes yielded negative values: –5.62 and –4.42 kg 1,4-DB eq for the physical and chemical routes, respectively. The negative values obtained for the Human Toxicity (HT) impact category should be interpreted with caution. In the CML-IA methodology, negative results can arise when the modeled system removes a toxic compound (Lufenuron 50-EC) from the aqueous matrix, thereby reducing the potential toxicity burden associated with the treated water compared to the untreated water. However, this does not necessarily imply a net beneficial effect in absolute terms, as the characterization model may not fully capture the long-term fate and ecotoxicity of Lufenuron 50-EC or its transformation products. Furthermore, the negative values may also reflect methodological artifacts, such as missing characterization factors for certain substances or compensatory effects among inventory flows. Therefore, the HT results presented here should be considered indicative of a potential reduction in toxicity burden rather than a quantified environmental benefit. Future studies should complement LCA with endpoint toxicity assessments or use alternative characterization methods (e.g., USEtox) to validate these findings.

The MAE impact category presents high values in both cases, physical process (4592.10 kg 1,4-BD eq) and chemical (2806.86 kg 1,4-BD eq), indicating a critical point in the life cycle. Considering the similarity between both processes, it can be inferred that the impact is strongly influenced by a common stage or common input flow, probably energy consumption, activation or final disposal. Although the ODP values are small, the chemical activation process (3.42 × 10⁻⁷ kg CFC-11 eq) has a lower impact compared to the physical (9.13 × 10⁻⁷ kg CFC-11 eq). This can be attributed to a lower total electrical consumption (mainly in activation). Therefore, it indirectly presents a lower dependence on industrial systems where use and potential leakage of refrigerants are present. The PO impact category is presented as another comparable impact, although slightly higher for the physical activation process (0.02 kg C₂H₄ eq), which is attributable to secondary emissions from thermal processes. Finally, the TE impact category presents a similar behavior to the general trend of greater impacts from the physical route, with a value of 0.01 kg 1,4-BD eq versus 3.33 × 10⁻³ kg 1,4-BD eq for the chemical process, possibly due to the presence of higher solid waste during the physical process.

3.4. Contribution Analysis

The contribution analysis was performed for each life cycle stage across the ten impact categories. The following sections describe the key findings for each stage. Figure 2 presents the contribution (%) by impact category on the adsorption life cycle of Lufenuron 50-EC using OPS-derived activated carbon for both activation routes.

As shown in Figure 2c, for physical activation, the activation stage (light orange bars) dominates nine of ten impact categories, contributing 80.6% to GWP, 95.0% to ODP, and 90.9% to MAE. In contrast, for chemical activation (Figure 2g), the activation stage (green bars) contributes 54.4% to GWP and 38.2% to FWAE, while the adsorption stage (Figure 2j, dark gray bars) dominates with >70% contribution to GWP, AD, AC, FWAE, HT, MAE, and PO. Final disposal (Figure 2k, blue bars) accounts for 56–72% of impacts for Route 1, but only 28–44% for Route 2.

• Transportation:

Figure 2a presents the contribution (%) of environmental impact for each transportation impact factor related to the processes, more specifically to the collection and transportation of OPS. The results show the differential distribution of the impact between both sub-stages. It can be observed that the OPS collection process presents a higher percentage of contribution in most of the environmental impact categories associated with the ecosystem, including Eut (78.68%), FWAE (87.44%), MAE (65.42%), ODP (72.24%), and TE (77.07%). This suggests that the activities associated with the collection and initial handling of OPS generate significant environmental burdens, attributable to the targeted use of energy resources, collection machinery, associated emissions, or even prior treatments. In parallel, the transportation of raw materials shows a percentage predominance in the categories of GWP (51.67%), AD (82.24%), AC (85.86%), HT (81.43%) and PO (63.61%), consistent with the emissions associated with the combustion of fossil fuels during transportation.

• Grinding and sieving:

Figure 2b shows that the environmental impact attributable to raw material pretreatment processes, specifically grinding and sieving, is negligible, with contributions of less than 3% in all impact categories assessed, with low values such as those for Eut (0.05%) and ODP (0.12%). This marginal impact on the environmental profile from the grinding and sieving stage is favorable from the perspective of pretreatment sustainability. Additionally, the differential distribution of the impact between the OPS collection and raw material transportation substages is maintained.

• Devolatilization:

From an environmental point of view, the devolatilization stage represents a critical phase, since it transforms the lignocellulosic material (OPS) into biochar through an intensive thermal treatment. In Figure 2e it can be observed that this stage presents the highest contribution (%) in almost all impact categories, with notably high values in GWP (58.89%), AC (57.51%), FWAE (55.46%), MAE (88.13%), ODP (94.94%) and TE (71.18%), which suggests a high use of energy resources, probably associated with the use of non-renewable thermal sources. The raw material transport stage significantly impacts AD (43.45%), HT (43.71%) and PO (33.13%); these results can be associated with emissions during road transport of the raw material from the point of collection to the place of implementation. The OPS collection also presents significant impacts in categories such as Eut (70.43%) and FWAE (38.71%), possibly associated with the generation of organic waste and the use of agricultural machinery. Finally, the grinding and sieving stage cannot be considered an environmentally critical stage due to its low contribution (<1% in most of the impact categories assessed).

• KOH-impregnation:

The potassium hydroxide (KOH) impregnation stage of biochar, although considered a key stage in the chemical modification of the material, presents moderate environmental contributions (<14%), with a higher presence in HT (13.77%), as shown in Figure 2f. Of the stages preceding the impregnation process, devolatilization remains the most critical stage due to its high contributions (%) in 9 of the 10 impact categories evaluated, surpassed only by the OPS collection in the Eut category (69.96%). The raw material transport stage significantly impacts several impact categories including AD (42.33%), AC (35.61%), HT (37.91%) and PO (32.37%).

• Activation:

The activation stage is, by a wide margin, the one that generates the greatest contribution to the environmental impact in almost all the impact categories evaluated. In Figure 2c, it is observed that the activation for physical activation (Route 1) has the greatest contribution in 9 of the 10 impact categories, mainly standing out in GWP (80.58%), FWAE (71.77%), HT (84.42%), MAE (90.96%), ODP (95.01%) and TE (74.60%), which suggests that this stage centralizes the use of energy resources and intensive processes possibly associated with the extreme thermal conditions required. Additionally, in the specific case of the GWP category, the high percentage of contribution can also be attributed to the carbon emissions that occur during the thermal process [49]. Similar results were reported in a previous work [50] in which the activation stage emerged as the main contributor in 6 of the 12 impact categories evaluated with values above 70% of the total contribution, mainly attributed to the energy demand of the thermal process. Similarly, Pereira et al. [27] reported that chemical activation with NaOH resulted in a more distributed impact profile across stages compared to physical activation, which is analogous to the behavior observed in this study for KOH-activated carbon. Also, it can be observed that the transportation of raw materials presents relevant contributions in AD (30.85%), AC (27.60%), PO (24.21%) and with a moderate contribution in GWP (9.96%), reflecting the environmental burden of this stage, possibly due to the use of fossil fuels. For its part, the collection of OPS stands out for presenting the highest contribution in the Eut category (63.82%), in addition to contributing significantly in FWAE (24.54%) and moderately in TE (19.53%) and PO (13.85%). This suggests that the environmental burden of this phase mainly affects water bodies, associated with the removal of organic waste or its final disposal. Grinding and sieving continue to present negligible contributions (<1%).

In the case of chemical activation (Route 2), it can be observed in Figure 2g that activation and devolatilization are the most critical stages from an environmental point of view. The activation stage presents the greatest impacts in the categories of GWP (54.41%), FWAE (38.22%), and HT (70.78%); this is attributed to the intensive energy use during the thermal treatment and chemical agents associated with KOH impregnation. These results align with those reported by [31]; they produced activated carbon from plastic waste by implementing different impregnation agents (NaOH, HCl, H₃PO₄, and KOH) for the removal of hazardous antibiotics, finding that KOH-activated carbon presents the greatest environmental impact, particularly in GWP. In parallel, devolatilization presents high contributions in categories such as AC (41.88%), MAE (64.42%), ODP (93.23%), and TE (61.29%). Although energy consumption in the devolatilization thermal treatment contributes to these high impacts, the largest contribution is strongly related to the use of nitrogen as an inert atmosphere during thermal treatment, since its production and consumption generate significant indirect emissions related to these impact categories. While the raw material transport and OPS collection stages are not dominant in most categories, they have a significant impact in AD (29.35%) and Eut (62.77%), respectively.

In comparative terms between Route 1 and Route 2, it is notable that the activation stage is positioned as a critical environmental point in both alternatives, albeit with different impact profiles. For Route 1, activation represents the largest contribution in 9 of the 10 impact categories evaluated, with the exception of the Eut category, which is led by the OPS collection stage. These high contributions are mainly associated with the high energy consumption of the muffle furnace. Meanwhile, although the activation stage for Route 2 dominates in categories such as GWP, FWAE, and HT, its values are more dispersed due to the significant participation of the devolatilization stage, which is mainly driven by the use of gaseous nitrogen.

Furthermore, while in physical activation the generation of impacts is centralized in the intensive use of energy to reach the temperatures required in thermal processes, in chemical activation the impacts are diversified between the thermal process, KOH-impregnation, and the use of nitrogen, introducing new vectors of environmental impact.

• Washing with distilled water:

A stage specific to Route 2, this is necessary to remove traces of the impregnating agent (KOH) deposited on the surface of the activated carbon. As seen in Figure 2h, the washing stage, although to a lesser extent, has significant impacts on TE (22.10%) and ODP (19.98%), suggesting the need to consider alternatives with lower environmental demands that help optimize this stage. The devolatilization and activation stages continue to be critical points due to their high-percentage contributions to the impact categories evaluated. In contrast, previous stages such as KOH impregnation, raw material transport, and grinding and sieving reflect more dispersed and/or marginal contributions depending on the impact category. Transport is particularly notable in the AD, AC, and PO stages. The OPS collection stage makes the largest contribution in the Eut category (60.91%).

• Dried in oven:

The drying stage is similar to the grinding and sieving stages in terms of the environmental impact generated. In both cases, their contribution is practically negligible (<0.2%), as shown in Figure 2i. When compared with Figure 2h, no significant differences are observed, further highlighting the low environmental impact of the drying stage. A normalized contribution analysis was performed using OpenLCA v2.4. As shown in Supplementary Materials, the drying stage contributes <0.2% to most impact categories (GWP, AD, AC, Eut, FWAE, HT, MAE, ODP) and only reaches 1.09% for PO and 10.77% for TE during the adsorption stage. In all cases, the adsorption stage dominates the impact profile, contributing between 71% and 86% for most categories. Therefore, despite the high electricity consumption of the drying stage (72 kWh in the chemical route), its contribution to the overall environmental impact is negligible.

• Adsorption:

The adsorption stage shows the highest contribution to most impact categories for both routes. It is important to note that this contribution is cumulative: the adsorption stage consumes the activated carbon produced in previous stages (activation, devolatilization, impregnation), thereby inheriting all environmental burdens associated with its production. Therefore, the high contribution of the adsorption stage does not imply that the adsorption process itself is energy-intensive; rather, it reflects the accumulated impacts of the entire upstream chain plus the electricity consumed during adsorption (9600 kWh).

According to the analysis in Figure 2d, for Route 1 (physical), the adsorption stage contributes to 6 of the 10 impact categories evaluated, including GWP (48.59%), AD (48.36%), AC (43.73%), FWAE (54.16%), HT (69.79%), and PO (46.51%). Although activation dominates in categories such as ODP (91.12%), TE (53.92%), and MAE (51.43%), the adsorption stage remains a major contributor due to the accumulated upstream burdens.

For Route 2 (chemical), Figure 2j shows that adsorption presents a significantly dominant contribution in all impact categories evaluated with the exception of ODP, which is mostly influenced by the devolatilization step due to the use of nitrogen. The adsorption stage contributes more than 70% to GWP (76.34%), AD (75.09%), AC (71.60%), FWAE (79.04%), HT (86.52%), MAE (71.11%), and PO (74.12%).

The adsorption stage in this study was modeled based on laboratory-scale batch experiments, which involved mechanical stirring at 170 rpm for 20 h using a 50 mL glass reactor [22]. This configuration results in a high electricity consumption per volume of treated water (9600 kWh per m³ at laboratory scale, scaled linearly). It is important to recognize that industrial-scale adsorption processes typically employ continuous column reactors (fixed-bed or fluidized-bed), where energy consumption is dominated by pumping (typically 0.5–5 kWh per m³ of treated water), rather than mechanical stirring. Therefore, the energy demand reported here for the adsorption stage is not representative of industrial practice and significantly overestimates the environmental impact of this stage. The linear scaling approach further amplifies this overestimation, as laboratory batch processes do not benefit from the economies of scale and energy efficiencies of continuous industrial systems. Consequently, the absolute GWP, AD, and other energy-dependent impact values reported in Table 4 should be interpreted as laboratory-scale projections that are likely much higher than those achievable in real-world applications. A more realistic assessment would require modeling a continuous column adsorption system, which is proposed as a priority for future research.

• Final disposal:

Figure 2k shows that the final disposal stage of spent activated carbon presents significantly higher environmental contributions for Route 1 compared to Route 2 for all impact categories assessed, with values between 55.91% and 72.73%. This difference is attributed to the higher pollutant load accumulated throughout the life cycle in Route 1, primarily due to: (first) the higher electricity consumption during physical activation (79,240 kWh vs. 25,000–50,000 kWh for chemical activation), which generates greater emissions of greenhouse gases and other pollutants and (second) the absence of washing and drying stages in Route 1, which in Route 2 remove residual KOH and reduce the toxicity of the spent carbon. Consequently, the depleted activated carbon from Route 1 carries a higher environmental burden to the final disposal stage. This trend is consistent with the results presented in Table 4, where Route 1 shows higher impacts across all categories compared to Route 2.

3.5. Sensitivity Analysis

To evaluate the robustness of the LCA results and identify which parameters exert the greatest influence on the environmental outcomes, a univariate sensitivity analysis was performed. Three critical parameters were selected based on their contribution to the total impact and their inherent variability:

• Electricity consumption during activation:

This varied by ±20% relative to the base values (39.62 kWh for Route 1 and 25 kWh for Route 2). This range accounts for typical fluctuations in laboratory-scale furnace operation and potential measurement errors.

• Lufenuron 50-EC adsorption yield:

This varied by ±5% (absolute) from the base yields (93.54% for Route 1, 94.44% for Route 2). This range is based on the standard deviation observed among experimental replicates from our previous work [22].

• Transport distance of oil palm shells (OPSs):

This varied by ±50% from the base value of 39.6 t·km, representing alternative logistical scenarios (e.g., closer or more distant supply sources).

The Global Warming Potential (GWP) was selected as the reference impact category for sensitivity analysis, as it is the most widely reported and comparable metric in water treatment LCA studies. Results are summarized in

Table 5. Sensitivity analysis results: percentage change in Global Warming Potential (GWP) for selected parameters.

Parameter	Variation	GWP Change—Route1	GWP Change—Route1
Electricity consumption (activation)	−20%	−16.3%	−10.5%
	+20%	+16.3%	+10.5%
Lufenuron 50-EC adsorption yield	−5% (absolute)	+5.3% 1	+5.2% 1
	+5% (absolute)	−4.8% 1	−4.7% 1
Transport distance (OPS)	−50%	−5.0%	−4.8% 2
	+50%	+5.0%	+4.8% 2

¹ Calculated as GWP per gram of Lufenuron 50-EC removed, not per m³ of water. ² Route 2 shows similar sensitivity due to identical transport inventory.

.

As show in Table 5, the sensitivity analysis reveals that GWP is most sensitive to electricity consumption during activation, especially for Route 1 (±16.3% for ±20% electricity variation), while variations in transport distance have a marginal effect (<6%). The ranking of the two routes (chemical activation having lower GWP than physical activation) remains unchanged across all sensitivity scenarios, confirming the robustness of the main conclusion.

The sensitivity analysis reveals the following key insights:

• Electricity consumption in activation is the most influential parameter.

For Route 1 (physical activation), a ±20% variation in electricity consumption leads to a ±16.3% change in GWP. For Route 2 (chemical activation), the same variation results in a ±10.5% change. The higher sensitivity of Route 1 is attributed to the greater proportion of electricity-related burdens in the physical activation stage (80.58% contribution to GWP, Figure 2c) compared to the chemical route (54.41%, Figure 2g).

• Adsorption yield variations have a moderate effect when GWP is expressed per gram of pollutant removed.

A ±5% change in yield alters the GWP per gram of Lufenuron 50-EC removed by approximately ±5% for both routes (Table 5). However, when GWP is expressed per m³ of treated water (the functional unit of this study), the effect is smaller because the yield directly determines the mass of pollutant removed for the same water volume. This finding underscores the importance of selecting an appropriate functional unit that reflects the service provided (pollutant removal) rather than merely the volume treated.

• Transport distance has a marginal effect on GWP.

Even a ±50% variation in transport distance changes GWP by less than 6% for both routes. This result is consistent with the contribution analysis (Figure 2a,e), where transport accounted for less than 10% of GWP in the baseline scenario. Therefore, efforts to reduce the environmental footprint should prioritize energy-intensive stages (activation and adsorption) over logistics optimization.

• The relative ranking between routes remains unchanged.

Across all sensitivity scenarios evaluated (nine combinations in total, see Supplementary Materials, chemical activation (Route 2) consistently exhibits a lower GWP than physical activation (Route 1), with reductions ranging from 32% to 38%. This confirms that the main conclusion of this study—that chemical activation using KOH is environmentally preferable to physical activation for Lufenuron removal—is robust to reasonable variations in key parameters.

Limitations of the sensitivity analysis: The analysis presented here is univariate (one parameter varied at a time) and does not account for potential correlations between parameters (e.g., adsorption yield and activated carbon dose). A full multivariate uncertainty analysis (e.g., Monte Carlo simulation) was not performed due to the lack of probability distribution data for foreground parameters. This limitation is acknowledged, and future work should focus on collecting replicate data to enable probabilistic uncertainty assessment. Nonetheless, the univariate sensitivity analysis provides valuable insight into the relative importance of each parameter and confirms the qualitative robustness of the findings.

4. Conclusions

The adsorption stage is the main environmental hotspot for both activation routes. For chemical activation (Route 2), this stage contributes more than 70% of the burden in GWP, AD, FWAE, and HT, with a GWP of 75.86 kg CO₂ eq per m³ of treated water. For physical activation (Route 1), the adsorption stage dominates six of ten impact categories, with a GWP of 117.62 kg CO₂ eq per m³—approximately 55% higher than Route 2.

The higher impact of Route 1 is attributed to the integration of devolatilization within the activation step, which concentrates thermal energy demand. In Route 2, devolatilization is an independent step, reducing the relative fraction of impact attributed to thermal treatment.

Final disposal also shows higher contributions for Route 1 (56–72%) compared to Route 2 (28–44%), indicating greater accumulation of contaminants throughout the life cycle of the physical route.

To minimize impacts, the following strategies are proposed: optimize adsorption efficiency with renewable energy; optimize activation conditions and replace nitrogen; implement gas recovery or hydrothermal carbonization for devolatilization; and apply thermal regeneration or reuse depleted activated carbon.

Future research should explore optimized operating parameters, alternative impregnating agents with lower environmental burden, renewable energy integration, and depleted activated carbon regeneration. An economic analysis would complement the environmental assessment.

A limitation of this study is the simplified end-of-life assumption for depleted activated carbon, which was treated as a stabilized output without landfill burdens. This is justified by FTIR evidence of strong Lufenuron 50-EC-adsorbent interactions [22]; however, future research should include leaching tests and alternative EOL scenarios (e.g., thermal regeneration).

Another limitation of this study is the use of the European electricity mix from the ELCD 3.2 database as a proxy for the Colombian grid, as a specific Colombian electricity dataset was not available. Future research should develop a Colombia-specific electricity mix to improve regional accuracy and reduce uncertainty in the results.

A direct comparison of environmental impacts across different adsorbents for Lufenuron removal is not possible at this time due to the lack of published LCA studies for other materials. Future research should address this gap to enable a more comprehensive sustainability assessment.

A major limitation is the laboratory batch configuration, which consumes 9600 kWh per m³—two to three orders of magnitude higher than industrial continuous column systems (0.5–5 kWh/m³). Thus, energy-dependent impacts (e.g., GWP, AD) are substantially overestimated. Future work should develop a pilot or industrial-scale LCA model based on continuous column adsorption.