Multifunctional Performance of Bacterial Cellulose Membranes in Saline and Oily Emulsion Filtration

Abstract

The separation of oil-in-water emulsions from industrial wastewater remains a significant challenge, particularly under saline conditions. This study evaluated bacterial cellulose (BC) membranes from Komagataeibacter hansenii for filtering synthetic effluents with high oil content (ES1) and saline oil-in-water emulsions (ES2). FTIR confirmed the incorporation of lipophilic compounds into the BC matrix. Crystallinity decreased from 78.8% to 40% following ES1 filtration, while a new peak at 2θ ≈ 31.8° appeared in ES2, indicating salt deposition. TGA revealed increased mass loss in the oil-saturated membrane (BCO), whereas the saline-exposed membrane (BCOS) exhibited higher thermal stability. SEM showed fiber compaction and localized deposition of oil and salt, corroborated by EDS, which identified Na, Cl, Ca, and elevated oxygen levels. Mechanical testing indicated that oil acted as a plasticizer, increasing the elongation at break of BCO, while salt crystallization enhanced BCOS stiffness. The membranes removed up to 98% of organic load (BOD and COD), 69% of oils and greases, and reduced turbidity and apparent color by 92%. Partial salt retention (~23%) and a significant decrease in dissolved oxygen were also observed. These results demonstrate the potential of BC membranes as an effective and sustainable solution for the treatment of complex oily and saline wastewater.

1. Introduction

The discharge of oily pollutants into aquatic environments remains one of the most pressing environmental and operational challenges across multiple sectors of modern society [1]. Oil contamination arises from a wide range of sources, including large-scale incidents such as accidental spills during offshore petroleum extraction and maritime transport, as well as the continuous and often unregulated release of oily effluents from industrial operations, municipal wastewater treatment facilities, and domestic discharges [2,3]. The widespread and persistent nature of these sources makes oily pollution a complex and enduring environmental issue, with significant impacts on aquatic ecosystems and the quality of available water resources [4,5].

The adverse effects of oil contamination include habitat degradation, toxicity to aquatic organisms, bioaccumulation of hydrocarbons across trophic levels, and economic losses in sectors such as fisheries, tourism, and water supply [6,7]. The complexity of this issue is increased by the diverse physical and chemical forms that oils can take in aquatic systems, ranging from surface films to dispersed or emulsified phases [8,9]. Among these, oil-in-water and water-in-oil emulsions are especially challenging, as they form highly stable structures that resist separation. At the molecular level, these emulsions involve dynamic interactions between polar water molecules and non-polar organic compounds, often stabilized by amphiphilic surfactants that reduce interfacial tension and prevent coalescence [2]. Their stability is often enhanced by additives such as emulsifiers and antifoaming agents, as well as by mechanical agitation from environmental forces, such as wave action [10,11].

In saline wastewater, on the other hand, the presence of high concentrations of ions such as Na⁺ and Cl⁻ further modifies emulsion behavior by altering electrostatic interactions, disrupting hydration shells around surfactant molecules, and potentially compressing the electrical double layer. These ionic effects can increase emulsion stability or alter droplet aggregation and surface charge, making separation even more difficult [4].

To address these complex waste streams, various technologies have been developed and implemented, with integrated approaches combining physicochemical and mechanical separation processes commonly used. Conventional methods like coagulation-flocculation, flotation, and membrane filtration are widely employed; however, their efficiency can be limited under high oil loads or saline conditions. As a result, alternative materials and advanced separation systems are being developed to enhance performance, sustainability, and adaptability in treating stable oily emulsions [12,13].

The selection and performance of these treatment strategies depend heavily on the characteristics of the effluent, including the type of oil present, the degree of emulsification, and the concentration of suspended solids [2,4]. The limited efficiency of standalone methods, combined with the growing demand for treated water reuse in industrial circuits or compliance with environmental regulations, continues to drive the search for more effective, sustainable, and cost-efficient solutions [14,15].

In this context, increasing attention has been focused on the development of novel biotechnological materials that can enhance separation efficiency while minimizing secondary environmental impacts. Among these materials, bacterial cellulose (BC) has emerged as a promising alternative to conventional synthetic membranes in filtration systems [16,17]. Produced by microorganisms from genera such as Komagataeibacter and Agrobacterium, BC exhibits a three-dimensional nanofibrous structure, characterized by hydrophilicity, oleophobicity, and high mechanical strength —properties that make it particularly well-suited for removing oily contaminants from aqueous media [18].

The innovation of this work lies in the use of unmodified bacterial cellulose as a biosustainable filtration material that operates effectively even under saline and high-oil conditions. Unlike synthetic membranes, which often require chemical functionalization, BC naturally combines high porosity, nanometric fiber size, and amphiphilic behavior, enabling selective retention of oily compounds while allowing water to pass. The structure’s capacity to entangle oil droplets and partially retain salts results from both physical adsorption and mechanical interlocking between fibrils and emulsified contaminants.

Thus, the main objective of this study was to investigate the application of BC membranes for treating oil-in-water emulsions. For this purpose, the filtration performance was evaluated using effluents with varying oil concentrations and salinity levels, alongside an assessment of the treated water’s compliance with Brazilian environmental regulatory standards. Physicochemical characterization of the membranes was also conducted before and after filtration to identify structural changes associated with contaminant retention. The results provided insight into the potential of BC membranes as a viable material for treating oily effluents from various contamination sources.

2. Materials and Methods

2.1. Membranes Production

2.1.1. Microorganism and Culture Maintenance

BC was produced using Komagataeibacter hansenii UCP1619, a strain registered in the Culture Collection of the Environmental Science Research Center (NPCIAMB) at the Catholic University of Pernambuco (Brazil). The microorganism was cultivated in Hestrin–Schramm (HS) medium, adjusted to pH 6.0, containing 5.00 g/L yeast extract, 20.00 g/L glucose, 1.15 g/L citric acid, 2.00 g/L agar, 2.70 g/L disodium phosphate, and 5.00 g/L peptone. Incubation was carried out at 30 °C following the protocol established by Hestrin and Schramm [19]. Following this step, the culture was scaled and processed under defined conditions to enable cellulose production and downstream treatment.

2.1.2. Culture Conditions, Purification, and Production Yield of the BC

BC production was initiated by preparing a 100 mL pre-inoculum to promote microbial adaptation to the HS liquid medium. The pre-inoculum was maintained under static conditions at 30 °C for 48 h. Subsequently, the main culture was established by inoculating 3% v/v of the pre-inoculum into rectangular glass containers, each containing 800 mL of standard HS medium. The cultures were incubated statically at 30 °C for 14 days.

2.1.3. Membrane Recovery and Preparation for Filtration

Following the incubation period, the membranes were thoroughly rinsed with tap water to remove residual sugars and culture residues, then transferred to a beaker containing deionized water. For sterilization, the membranes were subjected to flowing steam in an autoclave at 121 °C for 10 min, a procedure repeated three times to ensure the removal of residual microbial cells and any trace components from the culture medium.

After this step, the membranes were stored in deionized water at 4 °C until use. Before each filtration experiment, they were gently blotted with absorbent paper, without rubbing or damaging the membrane structure.to remove excess water and immediately placed into the Büchner funnel setup without any further chemical or structural modification.

Finally, the membranes were weighed, and production yield (g/L) was calculated according to Equation (1).

$$AverageYield={\displaystyle \frac{Meanmembranemass\left(\mathrm{g}\right)}{MeanVolume\left(\mathrm{L}\right)}}$$

(1)

2.2. Preparation of Synthetic Effluents

Two different synthetic effluents (Figure 1) were prepared for the filtration experiments, with each effluent processed using a dedicated BC membrane. Effluent 1 (ES1) was designed to replicate water with critical levels of oil contamination. It consisted of DI water containing 70 mg/L of Fuel Oil (OCB2), a low-sulfur, high-viscosity fuel oil, and 1 mg/L of the chemical surfactant sodium lauryl sulfate to promote partial emulsification of the oil molecules.

Effluent 2 (ES2) was formulated to simulate an oil pollution scenario in a coastal or marine environment. It was composed of DI water with 35 g/L of sea salt, corresponding to the average salinity of seawater, dissolved with the aid of a heat source, and 17 mg/L of Fuel Oil (OCB2). To promote emulsification, both mixtures were subjected to gyroscopic agitation in a paint mixer at 250 RPM for 30 min, simulating the mechanical forces typically exerted by wave motion.

It is important to note that the effluents ES1 and ES2 were designed to simulate different real-world contamination scenarios. ES1 represents industrial discharges, where oil is often released at high concentrations with slight salinity. In contrast, ES2 mimics marine oil spills, which typically involve lower oil content dispersed in a highly saline environment.

2.3. Filtration Process

Given that the objective of this study was to evaluate the feasibility and efficiency of treating synthetic effluents using bacterial cellulose membranes, the experiments were conducted at a laboratory scale. The filtration system consisted of a 1000 mL filtering flask, a Büchner funnel, and a vacuum pump. The membranes were adapted to the system by direct marking with the funnel and trimmed to match the equipment dimensions.

Due to the variation in oil content between the different effluents, membrane fouling during operation had to be taken into consideration. Therefore, a fixed volume of 100 mL was selected for filtration. This volume was sufficient to enable subsequent physicochemical and microbiological analyses of the treated water, while also inducing surface fouling and partial pore sealing of the polymeric matrix.

2.4. Characterization of Bacterial Cellulose Membranes

2.4.1. Moisture Management and Hydrophilic Properties

The hydration behavior of BC membranes was characterized using a set of analytical techniques that targeted water retention, surface wettability, absorption kinetics, and structural swelling. WRC was determined by weighing samples before and after drying at 50 °C to constant mass, ensuring complete removal of residual moisture. The WRC (%) was calculated according to Equation (2):

$$WRC\left(\mathrm{\%}\right)={\displaystyle \frac{Meanwetweight\left(\mathrm{g}\right)-Meandryweigth\left(\mathrm{g}\right)}{Meanwetweight\left(\mathrm{g}\right)}}\times 100$$

(2)

For surface wettability analysis, rectangular samples (10 × 5 mm) were used for water contact angle measurements via the sessile drop method. A droplet of DI water was carefully deposited onto the membrane surface, and the contact angle was recorded one second after deposition using a goniometer equipped with a mirrorless digital camera (XT10, Fujifilm, Fuji, Japan) [20]. The drop was then monitored for 10 min, and the average time for complete absorption was recorded to estimate the sorption index [21].

The swelling ratio (SR) was determined by first drying the samples at 50 °C to a constant weight. The dried membranes were then immersed in distilled water at 25 °C for 24 h. After gently removing surface water with absorbent paper, the swollen samples were weighed, and SR (%) was calculated according to Equation (3):

$$SR\left(\mathrm{\%}\right)={\displaystyle \frac{Swollenweight\left(\mathrm{g}\right)-Initialweight\left(\mathrm{g}\right)}{Initialweight\left(\mathrm{g}\right)}}\times 100$$

(3)

2.4.2. Fourier Transform Infrared Spectroscopy (FTIR)

This technique is based on molecular vibrations that occur when a sample interacts with electromagnetic radiation in the infrared region. It is widely used for identifying functional groups present in the organic structure of the analyzed materials [22]. The resulting data produce a graph known as an interferogram, in which the detector response is plotted as a function of the optical path difference [23].

For the analysis, BC membranes previously dried in a convection oven were stored in a desiccator containing silica gel and maintained at 25 °C until spectral acquisition. Measurements were performed at room temperature in absorbance mode using an Alpha II spectrometer (Bruker, Karlsruhe, Germany) equipped with a diamond ATR accessory. Spectra were acquired from 32 accumulated scans at a resolution of 4 cm⁻¹, covering the spectral range from 4000 to 400 cm⁻¹.

2.4.3. Thermogravimetric Analysis (TGA)

The thermal stability of BC was evaluated using a TGA 2 Star System thermogravimetric analyzer (Mettler Toledo, Greifensee, Switzerland). Approximately 5.4 mg of the dried sample was placed in alumina crucibles and subjected to controlled heating from 30 °C to 600 °C at a rate of 10 °C/min under a constant nitrogen flow of 50 mL/min. The maximum thermal degradation temperature (T_max) was determined based on the peak of the derivative thermogravimetric (DTG) curve [24,25].

2.4.4. X-Ray Diffractometry (XRD)

The crystallinity index (CI) and crystallite size of BC were analyzed using XRD with a D8 Advance ECO diffractometer (Bruker, Germany), operating at 45 kV and 25 mA with Cu Kα₁ radiation. Measurements were performed in reflection mode, with scanning over a 2θ range of 4° to 90° at a rate of 5° per minute [26,27].

The CI was calculated using the Segal method (1959) [28], which is commonly applied to the analysis of cellulose-based materials, as shown in Equation (4):

$$Crystallinity\left(CI\right)\left(\mathrm{\%}\right)={\displaystyle \frac{AC}{AC+Aam}}\times 100\%$$

(4)

where

Ac = area of the peaks corresponding to crystalline regions.

Aam = area of the halo corresponding to amorphous regions.

The peak areas used for CI calculation were determined according to the method proposed by Canevarolo Jr. [23]. Diffractogram analysis was performed using X’Pert HighScore Plus software (PANalytical, version 3.0, 2009) and OriginPro 9, which supported peak deconvolution and quantification of crystalline and amorphous regions.

2.4.5. Optical Microscopy

Analyses were performed using a binocular stereoscopic microscope (20× to 80× magnification) and a basic binocular achromatic optical microscope (Olen^®, ProWay Optics & Electronics Ltd., Ningbo, Zhejiang, China). The membranes were examined after filtration and drying steps to assess macroscopic features of the bacterial cellulose and the presence of retained material on the membrane surface.

2.4.6. Scanning Electron Microscopy (SEM) and Energy-Dispersive Spectroscopy (EDS)

Dried BC membranes were mounted onto copper metal stubs using carbon-based double-sided conductive adhesive tape. The samples were then sputter-coated with gold for 30 s using an SC-701 Quick Coater (Tokyo, Japan). Qualitative elemental characterization was performed by EDS using a detector coupled to an FEI INSPECT S50 scanning electron microscope (Midland, ON, Canada). EDS enables the identification of chemical elements present in the sample, but does not provide information on molecular compounds. The detection limit for this technique is approximately 0.1% by mass concentration [29,30].

2.4.7. Flexibility and Mechanical Testing

Flexibility was assessed using the method described by Chen et al. [31]. 100 consecutive folds were applied along the same line on selected samples. Mechanical performance was classified according to the number of folds the membrane could endure before rupture, using the following criteria: poor flexibility (<20 folds), fair (20–49), good (50–99), and excellent (≥100).

Tensile strength at break and maximum elongation were measured to characterize the mechanical behavior of the membranes before treatment. Five rectangular specimens (2.5 × 7.0 cm) were prepared. The tensile tests were carried out at ambient temperature (approximately 25 °C), using a universal testing machine (BIOPDI, São Carlos, Brazil) equipped with a 5-kN load cell. The testing procedure followed the ASTM D882 standard, with a crosshead speed of 50 mm/min [32].

2.5. Water Quality Analysis Before and After Filtration

The physicochemical and microbiological characterization of the synthetic effluents, both before and after filtration, was conducted in accordance with the protocols outlined in the Standard Methods for the Examination of Water and Wastewater, 24th Edition [33]. These analyses aimed to assess the effectiveness of the filtration treatment. The following parameters were evaluated: total oils and greases, turbidity, apparent color, total hardness, biochemical oxygen demand (BOD), chemical oxygen demand (COD), total alkalinity, and electrical conductivity.

3. Results and Discussion

3.1. Comparative Analysis of Membranes Before and After Filtration

3.1.1. Water Retention Capacity (WRC)

The high WRC observed in BC membranes is attributed to the combined effects of their hydrophilic surface chemistry and highly porous, three-dimensional fibrillar network [34]. Hydrophilicity enhances molecular-level water affinity, while the interconnected pore structure accommodates substantial water uptake [35]. This behavior was confirmed in the present study, with WRC values exceeding 97% (

Table 1. Yield and Water Retention Capacity (WRC) of Bacterial Cellulose Membranes.

Average Weight (g/L)		WRC (%)
Wet	4 61.73 ± 14.38	97.71 ± 0.49
Dry	10.57 ± 1.64

). The retained water plays a functional role during filtration, occupying the pore space and limiting oil infiltration into the matrix, thereby contributing to the apparent oleophobic response.

3.1.2. Determination of Water Contact Angle and Swelling Ratio

The filtration performance of BC membranes is strongly influenced by their wettability, which influences how fluids interact with the membrane surface during separation. The hydrophilic surface of the membrane allows preferential water permeation while rejecting oil-based compounds, enabling effective separation of oil-in-water emulsions. In this study, the water contact angle of 33.47 ± 1.51° confirmed the hydrophilic nature of the BC membranes, consistent with literature values, and reinforces their potential for oil-in-water emulsion separation [36,37].

Additionally, the time required for complete absorption of a water droplet on the membrane surface averaged 10.92 ± 1.28 s. This rapid uptake reflects the material’s high-water affinity and further confirms its hydrophilic character.

The membrane’s ability to rehydrate after complete drying was also evaluated, as this is a relevant parameter for applications involving reuse, extended storage, or unintended dehydration [38,39]. A rehydration ratio of 66.03 ± 3.89% was observed, indicating that the material can regain a significant portion of its water uptake capacity after complete drying (Figure 2), consistent with its low contact angle and rapid droplet absorption time.

3.1.3. Fourier Transform Infrared Spectroscopy (FTIR)

FTIR was employed to elucidate the chemical structure of bacterial cellulose (BC) before and after filtration of synthetic effluents, providing complementary insights to the structural data obtained from XRD and EDS. The transmittance spectrum of the untreated (as-synthesized) BC, presented in Figure 3, exhibits distinct absorption bands that confirm the preservation of both the structural integrity and functional groups characteristic of the cellulose matrix. The peak at 3340 cm⁻¹ corresponds to the O–H stretching vibration associated with hydrogen bonding, reflecting the material’s hydrophilic nature and its capacity for water interaction [40]. The band at 2895 cm⁻¹ is attributed to the asymmetric stretching of –CH₂ groups, characteristic of the polymeric backbone [41]. Additionally, the absorption observed at 1654 cm⁻¹ is likely related to the bending vibrations of adsorbed water molecules [42]. In the spectral region between 1500 and 1000 cm⁻¹, a series of well-resolved absorption bands characteristic of cellulose is evident. These correspond to vibrational modes associated with the glycosidic ring structure, including C=C stretching, C–H and O–H bending, as well as C=O and C–O–C stretching vibrations. A detailed assignment of these bands is provided in

Table 2. Characteristic vibrational bands and corresponding functional groups identified in BC membrane samples by FTIR.

Wavenumber (cm−1)	Spectral Assignment
3340	O–H stretching vibration
2895	Asymmetric C–H2 stretching vibration
1654	O–H bending vibration (absorbed water)
1547	C=C stretching vibration
1455–1428	C–H2 bending vibration
1362–1206	C–H bending vibration
1161–1108	C–O–C stretching vibration
1054–1000	C–O bending vibration

[43].

Following the filtration of the ES1 effluent, distinct changes were observed in the FTIR spectrum, characterized by more intense absorption peaks and reduced transmittance in specific bands (Figure 4). In addition to changes in transmittance at 3343 and 2854 cm⁻¹, a new absorption band emerged at 2922 cm⁻¹, attributed to the asymmetric stretching vibration of C–H groups. This feature suggests the incorporation of lipophilic residues into the BC matrix during the filtration process (

Table 3. Functional groups identified in bacterial cellulose membranes exposed to oil-containing effluent (BCO).

Wavenumber (cm−1)	Spectral Assignment
3343	O–H stretching vibration
2922–2854	Asymmetric C–H stretching (CH2)
1652	O–H bending
1548	C=C stretching
1457–1429	CH2 bending
1373–1206	CH bending
1161–1005	C–O stretching

). According to Balistreri et al. [41], increases in peak intensity, area, and sharpness are indicative of a higher density of chemical bonds, which is consistent with the fact that OCB2 contains functional groups similar to those found in native BC [14]. Furthermore, previous studies indicate that BC membranes are capable of assembling at the oil–water interface, promoting micelle stabilization and influencing the interactions between the polymer and surrounding components. This interfacial behavior may have promoted a reorganization of the hydrogen-bonding network within the BC structure [44].

A comparison between the FTIR spectra of membranes exposed to ES1 and ES2 revealed no significant differences, as both effluents contained the same oil (OCB2) as the primary contaminant. Sodium chloride (NaCl), a simple ionic compound, does not exhibit characteristic absorption bands in the mid-infrared region due to the absence of oscillating dipole moments in its ionic bonds [45].

As such, NaCl is commonly used as an infrared-transparent window material in FTIR spectroscopy. Any spectral features related to NaCl can typically only be observed indirectly, such as through its interactions with water molecules or organic compounds. Consequently, the direct detection of NaCl via FTIR is inherently limited [45]. Therefore, the spectral modifications observed are more directly attributed to the presence of oil in the system, which is likely responsible for the interactions with the polymer matrix. The corresponding peaks are illustrated in Figure 5 and listed in

Table 4. Chemical groups identified in the FTIR spectra of bacterial cellulose membranes with oil and salt (BCOS).

Wavenumber (cm−1)	Spectral Assignment
3347	O–H stretching vibration
2954–2854	Asymmetric stretching of CH2
1655	O–H bending vibration
1554	C=C stretching vibration
1458	CH2 bending vibration
1376–1207	CH bending vibration
1162–1012	C–O stretching vibration

.

Considering the spectral changes observed after filtration of oily effluents and the absence of significant structural alterations in the membrane despite the presence of surfactants or saline components, it can be concluded that BC demonstrates structural resilience under such conditions. This highlights the membrane’s substantial potential for treating industrial effluents containing synthetic surfactants and for use in high-salinity environments such as seawater, where it ensures effective separation of oil residues.

3.1.4. Thermogravimetric Analysis (TGA)

TGA was conducted to assess the thermal stability of BC membranes before and after the filtration of synthetic effluents containing oil (BCO) and oil with salt (BCOS). The curves shown in Figure 6 illustrate mass loss as a function of temperature under an inert atmosphere, revealing distinct thermal profiles among the samples.

The BC membrane (green), composed solely of purified cellulose, exhibited a thermal profile characteristic of cellulose-based materials. An initial mass loss of approximately 5% was observed up to 120 °C, attributed to the evaporation of physically adsorbed moisture. The main degradation phase occurred between 250 and 370 °C, with a significant mass loss of around 52%, associated with the cleavage of glycosidic bonds and the subsequent pyrolysis of the cellulose structure [46]. The final residual mass at 600 °C was 25.2%, consistent with its predominantly organic composition and the absence of inorganic components.

Thermal analysis of the BCO sample (blue), following filtration of fuel oil-containing effluent (70 mg/L), revealed a shifted degradation pattern. The main mass loss began at an earlier temperature, around 220 °C, resulting in a total weight loss of approximately 96.6%. The final residual mass was only 3.4%, highlighting the high volatility of the retained hydrocarbons. These compounds act as thermal plasticizers, increasing molecular mobility and facilitating the degradation of the cellulose matrix. According to Rutkowski and Kubacki [47], the presence of oil promotes a more pronounced and homogeneous decomposition, favoring complete pyrolysis of the composite material. This finding is particularly relevant for potential post-filtration applications involving thermal regeneration or energy recovery of membranes saturated with hydrocarbon-based contaminants.

In contrast, the BCOS sample (red), which filtered the effluent containing oil and 35 g/L of sea salt, retained high thermal stability throughout the heating process.

Degradation of the cellulose matrix occurred more gradually, with the main mass loss taking place between 260 °C and 400 °C. This behavior is attributed to the presence of NaCl, an inorganic compound with a high melting point (801 °C), which remains stable at elevated temperatures and contributes directly to the residual mass. Additionally, NaCl may promote the formation of carbonaceous residues by hindering the diffusion of volatile degradation products, thereby limiting their release during pyrolysis [48,49]. This synergistic effect between oil and salt results in a membrane that is more resistant to thermal degradation, yet also more persistent in terms of solid residue accumulation.

The low residual mass of the BCO sample indicates potential for efficient thermal disposal, whereas the high thermal stability of the BCOS sample underscores the need for specific treatment routes, such as controlled incineration or reuse in composite matrices [50].

Moreover, the multiple degradation stages in the BCO and BCOS curves indicate that contaminants significantly alter the structure and degradation kinetics of BC, possibly promoting the formation of intermediate or residual phases. These findings contribute to a deeper understanding of the saturation mechanisms and may inform the development of environmentally appropriate strategies for the regeneration or disposal of BC membranes used in contaminated industrial settings.

3.1.5. X-Ray Diffractometry (XRD)

This analysis revealed significant structural changes in the BC membranes following their use in the filtration of oily and saline emulsions. As shown in Figure 7, the pure BC sample exhibited well-defined diffraction peaks at 2θ ≈ 15.16°, 17.44°, and 23.35°, corresponding to the (100), (010), and (110) planes of cellulose Iα. These reflections confirm the high initial crystallinity of the membrane. The CI of 67.36% further confirms the organized structure of the pure membrane, which is consistent with values typically observed in BC samples synthesized under homogeneous conditions and in the absence of contaminants [51].

In the sample filtered with ES1 effluent, the characteristic cellulose peaks showed a pronounced decrease in intensity, accompanied by the emergence of new reflections at 2θ ≈ 14.55°, 16.68°, 23.35°, 28.82°, and 31.45°, consistent with the incorporation of lipid fractions into the polymeric matrix. The 10% reduction in CI reflects a structural alteration within the crystalline domains. This change likely arises from interactions between polymer chains and bulky molecules, such as oil, which disrupt the packing of ordered regions and lead to a partial loss of structural organization [17]. Similarly, Medeiros et al. [15] demonstrated that reduced crystallinity compromises membrane hydraulic permeability, as contaminant accumulation within the pores decreases water affinity and hinders fluid transport through the membrane matrix.

The membrane used for filtering the ES2 effluent exhibited, in addition to cellulose and oil-associated peaks, a high-intensity peak at 2θ ≈ 31.8°, consistent with crystalline salts, likely chlorides or metallic carbonates, originating from the effluent itself [52]. The presence of these salts suggests that inorganic contaminants have deposited on the surface and within the pores of the membrane, potentially accelerating structural degradation during prolonged use. This structural alteration aligns with the findings of Barud et al. [51], who reported that the introduction of bulky or nonpolar groups hinders cellulose chain packing, increases interfibrillar distance, and results in less-defined diffraction patterns, an effect analogous to that induced by the adsorption of ions or hydrophobic molecules within the matrix.

The reduction in crystallinity, coupled with lipid retention and salt accumulation, substantially alters the membrane’s microstructure, impairing both water transport capacity and mechanical integrity. These combined effects compromise long-term functional performance, underscoring the importance of implementing periodic regeneration strategies or planned replacement to ensure efficiency and extend the service life of oil–water separation systems.

3.1.6. Optical Microscopy

Upon completion of the filtration process, the membranes were analyzed for signs of water loss, mechanical deformation, and retained contaminants. The membranes were then cut to specific dimensions for mechanical testing (7.0 × 2.5 cm) as shown in Figure 8 and taken for optical analysis.

Macroscopic evaluation of the “BC” sample revealed that drying occurred in a non-uniform manner. As moisture was lost, localized shrinkage occurred, leading to fiber bundling and the formation of elevated, darker regions across the membrane surface. These aggregated zones became increasingly evident under higher magnification, suggesting microstructural reorganization during dehydration (Figure 9).

The progression of the images reveals an initially homogeneous surface with slight porosity and subtle undulations, evolving into a more compact and wrinkled structure at higher magnifications. The presence of parallel lines and grooves indicates the superficial entanglement of cellulose nanofibrils, characteristic of the three-dimensional organization of bacterial cellulose (BC). The topography becomes more evident, with marked elevations and depressions formed during the drying process, highlighting the structural complexity of the membrane and its potential for applications in filtration and molecular load support [14,53].

In the BCO membrane (Figure 10), the effect of vacuum pressure is evident, as shown by the grooves matching the shape and dimensions of the Büchner funnel pores. Additionally, the membrane exhibits a significantly darker coloration, particularly within these deformed regions, suggesting the localized accumulation of contaminants driven by suction during filtration.

The region selected for higher magnification (Figure 10A) corresponds to an imprint left by one of the pores of the Büchner funnel on the BCO. In this region, the cellulose fibers appear more densely packed, likely due to the mechanical pressure exerted during the vacuum filtration process. In these regions, smaller oil molecules likely became trapped between cellulose fibers during filtration and, upon drying, acted as adhesive agents, reinforcing the local compaction caused by mechanical deformation and further reducing inter-fiber spacing [14].

In Figure 10B, a micro-tear in the membrane’s superficial layer is evident, likely caused by vacuum-induced mechanical stress during filtration. The torn segment remained compact during filtration, entrapping contaminants, and only became evident during the subsequent drying step. Figure 10C corroborates localized fouling: particulate material is confined within the interfibrillar network, and the more pronounced yellow discoloration indicates residual oil. Despite partial surface damage (see Figure 10B), the BC nanofibrillar architecture retains its selective retention capability via capillary confinement of hydrophobic species [54]. In addition, the localized yellowish discoloration observed in Figure 10C, more pronounced than in the preceding images, can serve as a qualitative marker of membrane saturation, supporting the need for regeneration or scheduled replacement after a defined number of operating cycles.

When examining the BCOS membrane (Figure 11), distinct patterns are observed. The principal effect reflects the effluent composition rather than intrinsic membrane properties. The emulsion’s high salinity, compounded by concentration during filtration and drying, drives NaCl crystallization and the accumulation of oil–salt deposits on the membrane [55].

Figure 11A shows whitish surface deposits consistent with crystalline NaCl aggregates concentrated at high-salinity sites on the membrane. Unlike BCO, the Büchner-imprint grooves are not evident, indicating that in situ salt crystallization with oil co-deposition dominates the post-filtration surface topography. These localized deposits induce internal stresses and compromise surface integrity [56,57,58]. High-magnification images (Figure 11C) confirm NaCl crystals with adherent oil forming localized fouling deposits, which explains the altered morphology after drying [58,59].

Oil–salt micro-aggregates appear to deposit within the cellulose fiber network, hindering uniform water evaporation and creating locally compacted, stiff regions. This heterogeneous fouling produces opaque, textured surface zones and can degrade performance in subsequent filtration cycles by altering porosity, flexibility, and rehydration capacity [60].

The micrographs show that BC membranes retain particles spanning diverse sizes, morphologies, and chemistries, even under vacuum-induced pressure. Superficial micro-tears associated with deformation at Büchner-pore imprints were observed, but no through-thickness rupture occurred. Oil, alone or combined with salts, promotes localized pore occlusion and fiber compaction without compromising overall structural integrity [61,62].

3.1.7. Scanning Electron Microscopy (SEM) and Energy Dispersive Spectroscopy (EDS)

To better observe the surface morphology of the membranes before and after the filtration process, scanning electron microscopy (SEM) was performed. This test enabled us to visualize how the presence of contaminants affected the fibers and the 3D spatial organization of the membrane, as well as how the permeability of the effluents penetrated the membrane fibers in different ways, thereby allowing us to trace a membrane saturation profile in the presence of various types of effluents.

As can be seen in Figure 12A–H, the lower the applied magnitude, the more uniform the membrane appears. However, at 200× magnification (B), we can already perceive the overlapping nature of the membrane fibers, with a rough, wave-like appearance [63].

The wide-field images (Figure 12E–H) confirm the nanometric profile of the bacterial cellulose membrane, with an average fiber diameter of approximately 87 nm. The non-uniform fiber formation, a result of the microorganisms’ metabolic activity and growth patterns, leads to the development of an interwoven three-dimensional network. This unique morphology promotes the creation of a porous mesh with high surface area, enabling efficient retention of larger molecular contaminants such as oils and greases, even under flow and pressure conditions typical of filtration systems [15,64].

Microscopy combined with energy-dispersive X-ray spectroscopy (EDS) analysis (

Table 5. EDS results of bacterial cellulose membrane before filtration process.

Element	Mass (%)
	A	B	C	D	E	F
C	48.76	48.29	48.93	49.12	48.44	48.67
O	48.53	48.97	48.21	48.11	48.88	48.45
S	1.62	1.09	0.50	0.68	2.05	1.99
Mg	0.87	0.83	1.54	0.18	0.50	0.42
P	0.22	0.82	0.82	1.91	0.13	0.47

) revealed that, although the membranes underwent a purification process, residual traces of elements from the original culture medium remain embedded within the cellulose matrix. The composition was predominantly carbon and oxygen, as expected for bacterial cellulose. However, low concentrations of sulfur, magnesium, and phosphorus were detected, indicating residual compounds of the culture medium [14].

Although these concentrations are minor, the residual presence of such elements may subtly influence membrane performance, particularly under critical filtration conditions. Phosphorus, magnesium, and sulfur may serve as anchoring points for salts or polar molecules, promoting localized accumulation and potentially contributing to pore blockage over time. While they are not the leading cause of membrane saturation, these residues underscore the importance of strict purification protocols, particularly for applications that require high selectivity or extended reuse cycles [65,66].

Upon analyzing the micrographs of the BCO membranes, significant alterations in surface morphology become evident, especially in the lower magnification images (Figure 13A,B), when compared to the pure BC membranes. The BCO surface appears notably smoother, exhibiting fewer protrusions and irregularities across various regions. This more homogeneous appearance is accompanied by the presence of granules of differing sizes, irregularly distributed. Such uniformity is likely the result of mechanical pressure applied during the filtration process, which compresses the bacterial cellulose fibers. This compression is further enhanced by the presence of oil in the effluent, which, once retained between the fibers, acts as a natural adhesive. The lipophilic molecules infiltrate the porous structure of the membrane, facilitating a stronger entanglement of the nanofibrils and thereby reinforcing the cohesion between pressure-affected regions [65,67].

The observed granules of varying sizes correspond to oil particles retained after the internal saturation of the membrane. Once the internal porosity is filled, excess oil begins to accumulate on the external surface. This accumulation is a direct consequence of the composition of the ES1 effluent, which contains a mixture of fuel oil and surfactant, resulting in partially emulsified structures that are difficult to separate. The granules visible on the surface are the result of coalescence of these oil particles, forming clusters of different sizes depending on local emulsification density and the extent of oil penetration into the fibrillar network [14].

In the intermediate magnification images (Figure 13C,D), the morphological profile of the retained contaminant becomes more evident. A heterogeneous distribution of oily material is observed across the surface, with some areas showing higher accumulation while others remain relatively clean. This irregularity may be attributed to local variations in porosity or the interaction of the effluent flow with the membrane’s three-dimensional structure [67].

In the high-magnification images (Figure 13E,F), the entangled fibrillar arrangement becomes more pronounced, clearly revealing how the combination of filtration pressure and oil content resulted in denser compaction of the cellulose fibers. These fibers appear to be pressed tightly together, forming a more compact network in which the contaminant is interwoven and embedded within the membrane structure. This observation supports the hypothesis that oil, in addition to acting as a contaminant, also physically modifies the membrane’s topography, directly affecting its surface configuration and potential for reuse [65].

EDS analysis after the filtration of Synthetic Effluent 1 (ES1) revealed the presence of residual elements on the bacterial cellulose membranes (

Table 6. EDS results of the bacterial cellulose membrane after the filtration process of the ES1.

Element	Mass (%)
	A	B	C	D	E	F
C	46.57	47.98	47.01	47.85	45.13	46.74
O	52.11	51.02	51.75	51.01	53.96	52.11
Mg	0.11	0.25	0.19	0.20	0.12	0.28
Si	0.27	0.19	0.31	0.32	0.17	0.20
S	0.40	0.25	0.29	0.22	0.29	0.27
P	0.24	0.18	0.16	0.15	0.18	0.25
Ca	0.30	0.13	0.29	0.25	0.15	0.15

). These elements are not intrinsic to cellulose and were incorporated from the effluent contaminants, particularly surfactants, salts, and components of the fuel oil. Their detection reinforces the membrane’s ability to act as a selective barrier, retaining unwanted ionic and molecular species even after drying [68].

Elemental ratios also indicated a consistently higher oxygen content compared to carbon across all samples. This can be explained by the natural composition of cellulose, which contains hydroxyl groups, as well as the accumulation of oxygen-rich compounds on the membrane surface following filtration of oil-based emulsions [69].

The adsorption of fuel oil (OCB2) promotes the retention of compounds rich in oxygenated groups, such as esters, alcohols, and carboxylates. These molecules, combined with mild surface oxidation from air exposure and thermal drying, increase the detectable oxygen signal in EDS, which probes only the upper surface layer (up to 2 µm). Furthermore, as EDS does not detect hydrogen—an abundant element in the cellulose backbone—the apparent oxygen-to-carbon ratio becomes skewed. The O > C trend observed reflects the surface modification of the membrane caused by interactions with oil and surfactant residues from the effluent [70,71].

When analyzing the membranes used to filter the synthetic effluent ES2, even more pronounced structural differences become evident compared to the other samples. The high salt concentration in the effluent directly interfered with the scanning electron microscopy (SEM) imaging process, requiring a second round of metal coating to improve contrast and surface resolution. Despite this additional step, it was not possible to achieve magnifications beyond 1000×, likely due to the increased conductivity and charging effects caused by the accumulation of crystalline salts on the sample [72].

The presence of NaCl significantly altered the surface texture of the membrane. The resulting structure appeared rough, brittle, and uneven, as seen in Figure 14. This is most likely due to the infiltration and retention of NaCl crystals between the cellulose fibers, which, after the drying process, remained embedded and formed raised crystalline structures. These formations disrupted the otherwise smooth topography of the membrane, leading to a rugged relief. Such surface irregularities not only modify the physical appearance of the membrane but may also influence its mechanical behavior and reduce its potential for reuse in subsequent filtration cycles.

In the images obtained at 400× and 1000× magnifications (Figure 14C,D), the crystals present on the membrane surface become clearly visible. Their shapes appear distinctly irregular, with asymmetric contours and an absence of the sharp edges typically associated with crystalline salts such as sodium chloride (NaCl). Rather than displaying well-defined geometric features, these crystals are covered by a denser, amorphous layer of material, suggesting partial encapsulation by oily compounds. The interaction between salt and oil, both present in the effluent, likely results in a surface coating that obscures the usual morphological characteristics of the crystals, making their direct identification under electron microscopy more difficult [73].

This lipophilic coating directly interferes with the visualization of crystalline features at high magnifications, giving the surface a more diffuse and rough appearance. Furthermore, the coalescent behavior of the oil may contribute to the formation of heterogeneous aggregates, where different sizes and densities of material overlap. This complex configuration supports the hypothesis that the residual oil retained in the membrane acts not only as a contaminant but also as a surface-modifying agent, altering the way crystals organize and interact with the cellulose matrix after the drying process [73].

After the filtration of ES2, the EDS analysis revealed notable changes in the elemental composition of the sample. Although carbon (C) and oxygen (O) remained the major constituents, their contents slightly decreased compared to the pure membrane. This reduction is attributed to the occupation of the cellulose network by other elements from the effluent, suggesting the incorporation of both organic and inorganic contaminants into the BC matrix during the filtration process.

The notable presence of sodium (Na) and chlorine (Cl), confirmed by EDS analysis (see

Table 7. EDS results of the bacterial cellulose membrane after the effluent filtration process 02.

Element	Mass (%)
	A	B	C	D	E	F
C	43.17	46.98	45.27	43.89	44.13	43.62
O	43.73	39.87	40.77	43.05	42.40	42.77
Na	6.30	5.95	6.39	5.96	6.49	6.34
Cl	5.91	6.46	6.42	6.13	6.12	6.32
Mg	0.11	0.16	0.17	0.06	0.08	0.10
Si	0.13	0.14	0.25	0.24	0.25	0.26
S	0.23	0.16	0.26	0.22	0.31	0.30
P	0.22	0.15	0.24	0.21	0.11	0.10
Ca	0.20	0.13	0.23	0.24	0.11	0.19

), indicates the retention of salts from the marine salt used in the effluent formulation. The significant capture of these ions shows the membrane’s effectiveness in trapping saline components, likely due to its high porosity and the combined effects of capillarity and surface adsorption.

This retention suggests that salt ions are partially immobilized within the porous matrix, especially in regions of lower fluid mobility. Although partial ion occupation of pores could theoretically reduce the available surface area for organic retention, the membrane still exhibited high removal efficiency for organic contaminants (up to 98% of BOD and COD), indicating that the structural and chemical properties of bacterial cellulose remained active and functional even in saline conditions. This reinforces the potential of BC membranes to operate effectively in complex effluents with both oily and saline loads.

Small amounts of magnesium, silicon, phosphorus, sulfur, and calcium were also detected. Although present in trace levels, these elements reflect the complexity of the membrane after filtration of the effluent. The non-uniform distribution of these residuals may be linked to the BC’s three-dimensional structure and the affinity of the contaminants for the cellulose’s functional groups.

3.1.8. Mechanical Testing

Flexible membranes accommodate bending and handling-induced strain and sustain greater tensile elongation without brittle fracture, reducing crack initiation and extending service life. Their compliance also improves conformal contact under sealing pressure, although in-plane compressive loads generally produce wrinkling rather than increased load-bearing [74]. During tensile testing, the specimens exhibited distinct failure behaviors, indicating that retained contaminants modified the mechanical response of the BC network (Figure 15).

Mechanical analyses of the bacterial cellulose (BC) membranes revealed that the presence and type of retained contaminants significantly influenced tensile strength and deformation behavior (Figure 16). The control membrane (BC), composed solely of pure cellulose, exhibited a clear transverse fracture, breaking entirely from one edge to the other, with an average tensile strength of 90.925 ± 13.817 N and an elongation at break of 17.175 ± 1.746%. This profile reflects a homogeneous and porous structure, rigid and free from plasticizing agents, which promotes uniform stress distribution across the matrix and results in complete, brittle failure [64,75].

In contrast, the BCO membrane, saturated with effluent containing only oil, demonstrated the highest elongation at break (70.894 ± 4.821%) but intermediate tensile strength (93.225 ± 12.441 N), and did not fully rupture during the test, having been automatically interrupted by the equipment. The surface of these samples appeared smoother, with fewer irregularities. This behavior suggests that residual oil acted as a plasticizer, enhancing internal mobility and delaying crack propagation. While this increased ductility, it also reduced the membrane’s structural resistance under increasing load [76,77].

The BCOS membrane, used for filtering effluent containing oil and 35 g/L of NaCl, showed the highest tensile strength (193.315 ± 14.975 N) with an intermediate elongation of 49.381 ± 5.106%. Despite the oil content, its rupture behavior was distinctly abrupt and concentrated, indicating a more compact and rigid matrix. This can be attributed to salt crystallization between the cellulose fibers, which may have created ionic bridges or points of stiffness, temporarily increasing mechanical resistance while reducing flexibility. The visual aspect of these samples was rougher and more brittle, consistent with the presence of residual salt crystals on the surface [56].

Therefore, the fracture patterns observed align closely with the quantitative data from the mechanical tests, reflecting the complex interaction between retained contaminants and the three-dimensional fibrillar network of the membranes. Oil acts as a natural plasticizer, increasing flexibility and deformation capacity, whereas salt promotes rigidity and compactness, enhancing tensile strength at the expense of ductility. The chemical nature of the adsorbed contaminants has a direct impact on the structural performance of the membranes, particularly in applications that require prolonged mechanical loading or reuse over multiple filtration cycles [78].

3.2. Comparative Analysis of Water Before and After Filtration

Reductions in visible oil, apparent color, and turbidity were observed after filtration, consistent with the quantitative results in Figure 17 and

Table 8. Physicochemical characterization of synthetic effluents pre- and post-filtration by the BC membrane. Note: “<” and “>” indicate values below or above the reporting limit, respectively.

Parameter	ES1 (Before Filtration)	ES1 (After Filtration)	ES2 (Before Filtration)	ES2 (After Filtration)
pH	8.14	6.46	7.67	7.24
Dissolved oxygen (mg/L)	2.9	0.7	2.3	0.5
Total oils and greases (mg/L)	70.0	21.4	17.2	<10.0
Turbidity (NTU)	14.7	4.9	27.8	5.64
Apparent color (HU)	>500	123	233	18
Total hardness (mg/L as CaCO3)	950	119	1145	169
BOD5 (mg/L)	276	<5.0	241	<5
COD (mg/L)	762	<15	593	<15
Electrical conductivity (µS/cm)	<2	<2	73,000	56,000
Total alkalinity (mg/L as CaCO3)	792.4	83.9	327.7	95.8

.

For ES1, total oils and greases decreased from 70.0 to 21.4 mg/L (≈69% removal), consistent with selective water permeation through the hydrophilic fibrillar network and interfacial rejection of dispersed oil [79]. Electrical conductivity remained below 2 µS/cm before and after treatment, in line with the non-ionic, hydrophobic character of the fuel oil constituents that do not contribute to bulk conductivity [80,81]. Turbidity and apparent color fell from 14.7 to 4.9 NTU and from >500 to 123 HU, respectively, corroborating visual clarification.

Organic load was strongly reduced (COD 762.0 to <15.0 mg/L; BOD₅ 276.0 to <5.0 mg/L) [14,82]. Total hardness declined from 950.0 to 119.0 mg/L as CaCO₃ and total alkalinity from 792.4 to 83.9 mg/L as CaCO₃, indicating partial removal of mineral species and colloidal solids that interact with or are captured by the membrane. Dissolved oxygen decreased from 2.9 to 0.7 mg/L, which likely reflects residual oxygen demand, reduced gas–liquid exchange in the presence of oil films, and degassing effects during vacuum filtration [83,84]. Small apolar fractions in stable emulsions may still permeate the membrane, which explains the nonzero oil residual.

Under saline conditions (ES2), the membrane again achieved effective oil removal, with total oils and greases decreasing from 17.2 mg/L to <10.0 mg/L. Notably, this performance was sustained under high salinity, a condition that generally complicates oil–water separation [14]. Turbidity and apparent color decreased from 27.8 to 5.64 NTU and from 233.0 to 18.0 HU, resulting in near-clear filtrates, similar to those of DI water [85].

Organic load was strongly reduced (COD 593.0 to <15.0 mg/L; BOD₅ 241.0 to <5.0 mg/L) [86,87]. Electrical conductivity declined from 73 to 56 Ms/cm (≈23%), confirming only partial salt retention, which is consistent with ion passage through a hydrophilic micro-/nanofibrillar network lacking ion-exchange functionality [86,87]. The observed reduction likely arises from co-deposition of salts in the fouling layer and pore-scale occlusion by salt aggregates, in agreement with the FTIR/XRD evidence of crystalline deposits and the microscopy observations of oil-coated NaCl on the membrane surface. Total hardness and alkalinity also decreased (1145.0 to 169.0 mg/L as CaCO₃; 327.7 to 95.8 mg/L as CaCO₃), indicating partial mineral capture within the fouling layer and interfibrillar spaces. Dissolved oxygen decreased (2.3 to 0.5 mg/L), consistent with residual oxygen demand and inhibited gas–liquid transfer across oil films and salt-fouled surfaces; vacuum filtration can also lower DO via degassing [84,88,89,90].

4. Conclusions

The results obtained in this study demonstrated that bacterial cellulose (BC) membranes are highly effective in treating synthetic oily effluents, even under critical conditions involving high oil concentrations and elevated salinity. The membranes demonstrated a remarkable capacity for contaminant removal, achieving reductions of up to 98% in organic load (BOD and COD), 69% in oils and greases, and significant decreases in turbidity, color, hardness, and total alkalinity. Additionally, partial retention of salts was verified, with a 23.3% reduction in conductivity in saline effluents.

The structural and chemical analyses (FTIR, XRD, TGA, SEM/EDS) confirmed the presence of retained oil and salt components in the membrane matrix, as well as modifications in crystallinity, porosity, and surface morphology after filtration. The interaction between the contaminants and the cellulose network altered the mechanical performance of the membranes: the oil acted as a plasticizer, increasing elongation, while salt contributed to increased tensile strength due to its rigidifying effect. Despite these changes, the membranes retained their structural integrity and functionality, even after undergoing complete drying and rehydration cycles.

The filtration system developed in this work surpassed conventional methods in terms of efficiency and material sustainability. However, further studies are necessary to optimize membrane regeneration, improve flow dynamics under real-scale conditions, and ensure consistent performance in continuous or cyclic operations, aiming toward the successful scale-up and industrial application of BC-based filtration systems.