Advances in Graphene Oxide-Based Composites and Membranes: Structural Engineering, Multifunctional Performance, and Emerging Applications

Abstract

Graphene oxide (GO), with its high surface area, tunable chemistry, and exceptional mechanical, thermal, and electrical properties, is rapidly advancing as a transformative material in both composite engineering and membrane technology. In composite systems, GO serves as a multifunctional reinforcement, significantly improving strength, stiffness, thermal stability, and conductivity when integrated into polymeric, ceramic, or metallic matrices. These enhancements are enabling high-performance solutions across electronics, aerospace, automotive, and construction sectors, where lightweight yet durable materials are in demand. In addition, GO-based membranes are revolutionizing water purification, desalination, and other high-end separation technologies. The layered structure, adjustable interlayer spacing, and abundant oxygen-containing functional groups of GO allow precise control over permeability and selectivity, enabling efficient transport of desired molecules while blocking contaminants. Tailoring GO morphology and surface chemistry offers a pathway to optimized membrane performance for both industrial and environmental applications. This paper gives a comprehensive overview of the latest developments in GO-based composites and membranes, highlighting the interplay between structure, morphology, and functionality. Future research directions toward scalable fabrication, performance optimization, and integration into sustainable technologies are discussed, underscoring GO’s pivotal role in shaping next-generation advanced materials.

1. Introduction

Membrane science and technology play a crucial role in both chemical and environmental industries, offering diverse applications across separation processes [1]. A membrane can be defined as a selective barrier containing pores or channels that enable the controlled separation of molecules or ions from complex mixtures. The two key performance indicators of a membrane are permeability and selectivity, which are intrinsic material properties rather than characteristics of the fabricated membrane itself. For practical use, membrane materials must be engineered into suitable configurations—such as thin-film composite (TFC) or asymmetric membranes—consisting of an ultrathin selective layer supported by a mechanically robust, porous substrate. This combination ensures high permeance and selectivity while maintaining structural integrity. To achieve large-scale, efficient separation, membranes are typically assembled into plate-and-frame, hollow fiber, or spiral-wound modules that maximize surface area and flux. Understanding these fundamental principles provides essential guidance for researchers developing novel membrane materials and fabrication methods. It also clarifies why certain materials exhibiting excellent intrinsic permeability and selectivity as dense films may not be suitable for practical membrane module fabrication and operation [2].

At present, polymeric membranes dominate the global membrane market for applications such as water purification, desalination, and gas/vapor separations, owing to their operational simplicity, modular design, low cost, and energy efficiency. Despite these advantages, polymeric membranes face inherent performance limitations, known as the upper bound, which arises from the trade-off between permeability and selectivity [3,4]. In general, highly permeable polymers exhibit low selectivity, whereas highly selective polymers tend to have reduced permeability. To surpass this constraint, researchers continue to explore novel membrane materials that can achieve enhanced separation performance while maintaining processability into large-area membranes with ultrathin selective layers and porous mechanical supports. Inorganic membranes—such as those based on carbon, silica, and zeolites—have demonstrated superior intrinsic separation properties compared to conventional polymers. However, large-scale fabrication of defect-free inorganic membranes remains challenging. Similarly, only a limited number of polymers have reached practical application due to similar fabrication and stability issues. Consequently, the development of next-generation membrane materials capable of combining high selectivity, permeability, and scalability remains a critical focus for achieving energy-efficient and sustainable separation technologies (Figure 1).

In this context, graphene has emerged as a highly promising candidate for next-generation membrane materials, offering exceptional properties such as atomic-scale thickness (<1 nm), large scalable surface area, and remarkable tensile strength [6]. These features make graphene particularly attractive for achieving high membrane permeance, mechanical stability, and efficient separation performance, provided that suitable pores or transport channels can be precisely engineered. Nevertheless, pristine graphene is intrinsically impermeable to gases and ions because of the dense π–electron cloud within its aromatic structure, which prevents molecular penetration through the basal plane. Consequently, significant research efforts have been devoted to modifying graphene and its derivatives to introduce controlled defects or functional groups that enable selective transport.

Air pollution arising from industrial emissions of gases such as CO₂, CO, NO₂, and NH₃ represents a major environmental concern. Owing to the abundance of oxygen-containing functional groups on its basal planes and edges, GO can effectively interact with gas molecules through covalent and noncovalent bonding [7]. Consequently, GO has been widely investigated for gas adsorption, catalytic conversion, and storage [8]. GO-based membranes have demonstrated high CO₂ selectivity and permeability, while GO composites have shown remarkable adsorption capacities toward NH₃, acetone, formaldehyde, H₂S, SO₂, and NO_x. In addition, the unique electronic structure of GO supports photocatalytic CO₂ conversion into hydrocarbons or alcohols, providing a sustainable route for simultaneous carbon dioxide reduction and solar energy utilization [9,10,11]. Besides air purification, GO exhibits strong potential for water treatment applications. Its large surface area and rich surface functionalities enable the adsorption of heavy metal ions such as Pb²⁺, Cd²⁺, Co²⁺, and Au³⁺, as well as organic dyes including methylene blue and Rhodamine B. The incorporation of organic modifiers can further enhance its adsorption efficiency and selectivity [12,13,14,15,16,17]. GO exhibits strong potential in sensing applications due to its abundance of oxygen-containing functional groups, which enable selective interactions with target analytes through adsorption, redox activity, and surface binding. These tunable functional groups also enhance GO sensitivity by modulating electronic properties and facilitating signal transduction, making it a versatile platform for chemical and biosensing technologies [18]. Overall, the multifunctionality of GO—combining adsorption, catalysis, and photocatalysis—makes it a highly promising material for environmental remediation and pollution control.

The present review focuses on GO as a multifunctional material with remarkable potential across various technological domains. The discussion begins with an overview of GO synthesis methods and the resulting structural and chemical properties that define its performance in different applications. Emphasis is placed on how synthesis conditions influence GO morphology, degree of oxidation, and functional group distribution, which in turn determine its suitability for specific uses. The review further explores GO-based composite materials, highlighting the integration of GO with metals, metal oxides, and polymers to enhance mechanical, electrical, catalytic, and adsorption properties for diverse applications, including environmental remediation, energy storage, and catalysis. The final section of this review is devoted to GO membrane technology, covering the underlying principles and transport mechanisms that govern molecular and ionic separations. A detailed examination of fabrication techniques, types of GO membranes, including free-standing, supported, and GO-modified composites, and materials suitable for GO membrane construction is presented. By unifying these aspects, this work aims to provide a comprehensive perspective on the current state, challenges, and future prospects of GO and GO-based membrane materials for next-generation separation and environmental technologies.

Although several recent reviews have addressed graphene oxide-based composites or GO membranes, these works typically treat the two fields independently, without examining how advances in one domain inform progress in the other. The existing literature often focuses on synthesis, functionalization, or specific applications, but lacks a cohesive perspective that links GO’s structural evolution, composite engineering strategies, and membrane performance mechanisms. Furthermore, no comprehensive review currently integrates GO chemistry, composite formulation, and membrane fabrication into a single conceptual framework that highlights their interdependencies. By addressing these gaps, the present work offers a unified and comparative overview that bridges composite development and membrane science, thereby providing a clearer understanding of how GO’s intrinsic properties and modification pathways govern functionality across environmental, energy, and separation technologies.

2. Review Methodology

To ensure a comprehensive and transparent analysis of recent advances in GO-based composites and membrane technologies, a systematic literature search was conducted following a structured review methodology. The search covered a wide time period (more than 10 years) to cover most of the important research findings on the topic. Studies were retrieved from major scientific databases, including Scopus, ScienceDirect, SpringerLink, MDPI Journals, and Google Scholar, using different keywords, such as “graphene oxide” AND “composites”, “GO membrane” AND “water purification”, “GO–based nanocomposite” AND (“mechanical properties” OR “functional performance”), “graphene oxide” AND (“separation” OR “desalination” OR “filtration”), “GO” AND “structural engineering” AND “sustainable materials”. Additional keywords applied were as follows: GO modification, interlayer spacing, morphology tuning, multifunctional properties, antifouling, water remediation. Based on the Scopus database, entering keywords such as “graphene oxide” AND “composites”, approximately 37,756 documents were found across the ten-year time frame, with a diversity of subject areas (Figure 2, left), showing a substantial research effort devoted to the investigation of GO and its composites in different applications. Using keywords such as “GO membrane” AND “water purification”, 198 documents were found on the Scopus database, predominantly in the application areas of Materials Science, Chemistry, Chemical Engineering, Environmental Science, Engineering, Biochemistry, Genetics and Molecular Biology (Figure 2, right).

During the preliminary screening phase, some articles were excluded because they were off-topic, lacked relevance to GO composites or membranes, or did not address structural, functional, or targeted application-based aspects. A total of 350 full-text articles were subsequently assessed for eligibility. After applying the predefined inclusion and exclusion criteria—peer-reviewed publications, relevance to GO-based composites or membrane technologies, adequate methodological detail, and English language—a further 170 papers were excluded for insufficient data, unrelated material systems, absence of performance metrics, or limited structural analysis. Ultimately, 180 studies were considered for inclusion in the final review. Additional relevant articles identified through reference mining of key publications were screened qualitatively and incorporated when meeting the same criteria.

3. Graphene Oxide: Structure, Morphology, Synthesis and Application

Carbon-based nanomaterials represent a highly diverse class of materials distinguished by their allotropic versatility and tunable physicochemical properties arising from different carbon hybridizations (sp, sp², and sp³). Their classification is commonly based on dimensionality, which directly influences their electronic structure, surface area, and application potential (Figure 3) [19,20,21,22]. Zero-dimensional (0D) carbon nanomaterials, such as fullerenes and carbon quantum dots (CQDs), are characterized by quantum confinement in all spatial directions, leading to discrete energy levels, strong photoluminescence, and size-dependent optical and electronic properties. One-dimensional (1D) materials, including carbon nanotubes (CNTs) and carbon nanofibers, possess a high aspect ratio and anisotropic conductivity, rendering them valuable for nanoelectronics, composite reinforcement, and sensing applications. Two-dimensional (2D) materials, notably graphene, GO and reduced graphene oxide (rGO), exhibit outstanding charge carrier mobility, mechanical flexibility, and large specific surface area, making them promising for energy storage, catalysis, and electromagnetic interference (EMI) shielding. Three-dimensional (3D) carbon-based architectures, such as graphene foams, aerogels, and carbon networks, are assembled from lower-dimensional units to achieve interconnected porosity, mechanical robustness, and multidirectional conductivity. This classification framework underscores the direct relationship between dimensionality and functional performance, providing a foundation for tailoring carbon-based nanomaterials toward specific technological applications across electronics, energy systems, environmental remediation, and advanced composites [23,24,25,26,27]. In this review, the term ‘3D carbon materials’ refers not to fundamental carbon allotropes defined by IUPAC, but to three-dimensional engineered carbon architectures—including foams, aerogels, and monolithic networks—commonly used in carbon-based composite systems. This terminology is widely adopted in the materials science literature to describe macro- and mesostructured carbon frameworks that differ structurally and functionally from 0D, 1D, and 2D nanoscale carbons, yet remain highly relevant in the design of graphene oxide composites and membrane supports.

GO, a chemically modified form of graphene, has attracted widespread attention as one of the most versatile and processable 2D nanomaterials [28]. Unlike pristine graphene, which is composed entirely of sp²-hybridized carbon atoms arranged in a perfect honeycomb lattice, GO incorporates a wide variety of oxygen-containing functional groups, including hydroxyl and epoxy groups on the basal plane, as well as carbonyl and carboxyl groups at the sheet edges. This partial oxidation disrupts the extended π–conjugation network of graphene, thereby altering its electrical and mechanical properties while simultaneously introducing new physicochemical characteristics such as hydrophilicity, chemical tunability, and colloidal stability in aqueous and polar solvents. These attributes have made GO an indispensable precursor for graphene production and an independent material of interest across multiple fields, ranging from energy storage and environmental remediation to electronics, composites, and biomedicine [29].

The chemistry of graphene oxide is central to its functionality. The distribution and density of oxygen functionalities determine the extent of disruption in the sp² domains, which in turn govern its electrical conductivity, surface charge, and chemical reactivity. GO can engage in hydrogen bonding, electrostatic interactions, and covalent modifications, making it a highly adaptable platform for chemical functionalization. This tunability is particularly relevant in applications where surface chemistry dictates performance, such as in adsorption-based environmental technologies, catalysis, or biomedical interfaces where biocompatibility and surface anchoring are critical. Importantly, GO can also undergo reduction (chemically, thermally, or electrochemically) to form rGO, partially restoring its conductivity while retaining a degree of oxygen functionalities that allow for further chemical versatility. Thus, GO is not merely an oxidized form of graphene, but a chemically dynamic material capable of bridging between insulating, functionalized states and conductive, graphitic structures [30,31].

The structural and morphological aspects of GO are equally important. GO generally exists as single-layer or few-layer nanosheets, typically one atom thick yet spanning lateral dimensions from a few nanometers to several micrometers depending on synthesis and exfoliation conditions. These nanosheets exhibit a wrinkled and corrugated morphology, which arises from the presence of functional groups and structural defects, as well as from the inherent flexibility of the carbon lattice. Such morphology provides a high specific surface area and facilitates interactions with molecules, ions, or polymers in composite systems. Moreover, stacking and agglomeration tendencies are strongly influenced by the oxygen content and interlayer hydrogen bonding, which must be carefully managed when processing GO into films, coatings, or 3D architectures. The morphological diversity of GO thus enables a broad spectrum of material forms, ranging from colloidal suspensions to membranes, aerogels, and fibers [32,33,34].

In addition to the general structural features of GO, it is important to acknowledge the inherent variability in oxidation level, which directly governs the density and distribution of oxygen-containing functional groups. These functional groups not only influence GO’s hydrophilicity and chemical reactivity but also play a decisive role in composite formation by mediating interfacial bonding with polymers, metals, and metal oxides. Likewise, the lateral flake size of GO substantially affects its performance: larger sheets tend to promote more ordered laminar stacking and improved channel alignment in membrane architectures, whereas smaller flakes introduce additional edge defects that may enhance permeability but reduce structural uniformity. Recognizing these sources of morphological diversity is essential for understanding the broad spectrum of behaviors exhibited by GO-based composites and membranes.

The synthesis methods of graphene oxide critically determine its ultimate characteristics. The most widely adopted approach is the chemical oxidation and exfoliation of graphite using strong oxidizing agents, historically developed through Brodie, Staudenmaier, and Hummers’ methods. Among these, the modified Hummers’ method has become the most prevalent due to its relative efficiency and scalability. Studies show that modification in the graphite:KMnO₄ ratio can significantly influence the C:O ratio in GO samples [35]. Variations in synthesis route also led to significant differences in oxygen content, layer size, degree of exfoliation, and defect density, all of which influence GO’s performance in practical applications. For instance, harsher oxidation often yields highly functionalized but structurally degraded sheets, whereas milder methods can preserve larger sp² domains at the expense of lower oxygen content. In recent years, alternative approaches such as electrochemical exfoliation, microwave-assisted oxidation, and green synthesis routes using environmentally benign reagents have gained attention, reflecting the ongoing effort to tailor GO properties while improving sustainability [8,36].

Beyond traditional characterization, the interplay between GO chemistry, structure, and synthesis has become a central theme in understanding its performance. For instance, the hydrophilicity imparted by oxygen groups enables stable aqueous dispersions, which are essential for scalable processing into films or inks. At the same time, excessive oxygenation can severely compromise electrical conductivity, necessitating post-synthesis reduction strategies for applications in electronics or energy storage. Similarly, the lateral dimensions and defect density of GO sheets influence mechanical reinforcement in polymer nanocomposites, ion transport in membranes, and active surface area in catalytic or adsorption processes [37,38]. These structure–property correlations emphasize that GO should not be regarded as a single material, but rather as a diverse family of materials whose properties can be engineered through careful control of synthesis parameters and post-treatment strategies [32,39,40].

The scientific and technological importance of graphene oxide thus lies in its dual nature: it is both a functional material in its own right and a precursor for the production of reduced graphene oxide and graphene-based hybrids. Its oxygen functionalities provide sites for chemical anchoring and interactions, while its layered morphology ensures large surface areas and structural versatility. At the same time, its synthesis-dependent variability demands rigorous characterization and standardization, particularly as GO transitions from laboratory research to industrial applications. As such, a systematic understanding of its chemistry, structure, and synthesis routes is critical to rationally design GO-based systems with tailored performance.

In the following subsections, we will examine the structural and morphological features of GO, highlighting how nanoscale arrangement and defects shape its physical behavior. Next, we will explore in detail the chemistry of GO, focusing on the types and distributions of oxygen functionalities and their implications for reactivity and tunability. Finally, we will review the synthesis methods of GO, comparing conventional oxidation-based approaches with emerging techniques aimed at improving quality, scalability, and environmental sustainability. Collectively, this discussion will provide a comprehensive foundation for understanding the intrinsic characteristics of GO and their relevance in diverse applications.

3.1. Evolution of GO Structural Models

The structural understanding of GO has evolved significantly over the past century through a series of conceptual and experimental advancements (Figure 4). The earliest structural representation was proposed by Hofmann and Rudolf (1939) [41], who suggested that GO consists of graphite layers decorated with randomly distributed epoxy groups. Subsequently, Ruess (1947) [42] refined this model by incorporating hydroxyl functionalities and proposing an alternation between sp²- and sp³-hybridized carbon atoms within the basal plane, thereby introducing the first hybridized framework concept. Later, Scholz and Boehm (1969) [43] proposed a less ordered structural configuration characterized by C=C double bonds, periodically cleaved C–C linkages within corrugated carbon sheets, and the presence of hydroxyl and carbonyl groups in different chemical environments, notably without ether oxygen species. In 1994, Nakajima and Matsuo [44] introduced a stage–2 graphite intercalation compound (GIC)-like lattice model, based on the observation that fluorinated graphite oxide exhibited an X-ray diffraction pattern similar to that of stage–2 graphite fluoride (C2F).

A major milestone was achieved by Lerf and Klinowski (1998) [45] who utilized solid-state 13C and 1H nuclear magnetic resonance (NMR) spectroscopy to identify distinct carbon environments. They assigned the resonance at 60 ppm to epoxide (1,2–ether) groups and the signal at 130 ppm to aromatic carbon atoms and conjugated double bonds. They further observed that hydroxylated carbons induce local distortions, producing partial wrinkling within the layers. Based on these findings, they proposed a semi-aromatic, partially oxidized structural model composed of alternating unoxidized benzene-like domains and aliphatic six-membered ring regions. This Lerf–Klinowski (L–K) model has since become the most widely accepted representation of moderately oxidized GO [46].

Despite its broad acceptance, early structural models could not fully explain certain chemical characteristics, particularly the planar acidity of GO, which has since been attributed to the presence of specific oxygen functionalities and defect sites. Building on this gap, Szabó et al. (2006) [47] revisited and modified the Scholz–Boehm framework using comprehensive analytical techniques including elemental analysis, TEM, XRD, DRIFTS, XPS, ESR, and NMR. Their revised model excluded carboxylic acid groups and described GO as comprising two interlinked domains: regions of trans-connected cyclohexyl species containing tertiary alcohols and 1,3–ethers, and corrugated domains dominated by keto-quinoidal structures.

Further insights into the complex structure of GO emerged with the work of Rourke et al. (2011) [48], who investigated the behavior of GO in basic media. They demonstrated that alkaline treatment leads to its decomposition into two distinct components: a mildly oxygenated graphene fraction and strongly oxidized, graphene-bound oxidative debris (OD). This finding prompted the proposal of a two-component OD–base-washed GO model, which diverged significantly from earlier single-phase structural representations and fundamentally reshaped the understanding of GO chemistry. Rourke and co-workers also reported the metastability of unwashed GO, resonating with prior observations of metastable GO films [49], though the underlying mechanisms driving the structural instability under external stimuli remained insufficiently explored. Subsequently, Dimiev et al. (2013) [50] revisited the structural interpretation of GO through acid–base titration and ion-exchange experiments, focusing on its surface acidity. They proposed a Dynamic Structural Model (DSM), which extended beyond the static Lerf–Klinowski (L–K) framework by describing GO as a dynamic, hydrated, and pH-responsive system involving the reversible evolution of oxygen functionalities and carbon frameworks. More recently, Liu et al. (2018) [51] provided experimental evidence of C–O bonding configurations located both at the edges and within the basal plane of GO, thereby reinforcing aspects of earlier structural models, particularly the L–K model. Among the various models proposed between 1939 and 2018, the Lerf–Klinowski model remains the most widely adopted due to its strong correlation with experimental observations and adaptability for subsequent refinements, such as those introduced in the Rourke–Wilson and Dimiev–Alemany–Tour models. A recent study by Taniguchi et al. shows that GO undergoes reversible epoxide ring opening in alkaline media and ring closing in acidic environments under ambient conditions [52]. This pH-responsive behavior is enabled by the stability of carbocation intermediates and rapid proton migration along the GO surface. Experimental evidence further indicated that, under continuous flux, the irreversible formation of epoxides in base-treated GO arises from the decomposition of vicinal diols generated during alkaline ring opening. Because basal-plane epoxides constitute a major fraction of GO’s functional groups, such reversible transformations are central to understanding GO’s chemical reactivity and performance. Notably, epoxide ring opening provides a plausible explanation for the pronounced photoluminescence quenching observed in GO dispersions under alkaline conditions.

Overall, while these evolving models have deepened structural and chemical understanding, the 2D layered architecture of GO continues to be universally recognized as its defining feature. This fundamental characteristic has established GO as a pivotal material in nanoscience and technology, catalyzing extensive research activity, particularly following the 2010 Nobel Prize in Physics awarded for the discovery of graphene [53].

3.2. Morphology and Chemistry of GO

Morphologically, GO typically consists of thin, wrinkled, and transparent sheets derived from the exfoliation of graphite oxide. These nanosheets are generally a few nanometers thick and can extend laterally over several micrometers, depending on the precursor graphite and exfoliation efficiency. The wrinkled or crumpled morphology originates from internal strain caused by lattice distortions and the presence of oxygen functional groups on the basal plane and edges. Such surface corrugations enhance the mechanical flexibility of GO sheets and prevent restacking during dispersion and composite formation [30,33,54]. The porosity of GO is largely influenced by the synthesis method, degree of oxidation, and subsequent thermal or chemical treatments. In its native form, GO is typically a defect-containing, layered material with nanoscale non-idealities. However, upon reduction or assembly into aerogels, foams, or membranes, a hierarchical porous network can develop. This network includes micro-, meso-, and macropores formed through inter-sheet voids and the removal of oxygen groups during thermal reduction. The generation of porosity enhances ion and molecule diffusion, which is beneficial for applications such as catalysis, adsorption, and electrochemical energy storage. Studies have shown that mild thermal reduction or hydrothermal treatment of GO leads to partial restoration of the conjugated π–system and the evolution of gases (CO₂, CO, and H₂O), which in turn introduce nanoscale pores into the structure. Additionally, templated synthesis or activation with chemical agents (e.g., KOH, ZnCl₂) can further increase the pore volume and connectivity, transforming GO-based materials into high-surface-area frameworks [55,56,57]. The specific surface area of GO strongly depends on its exfoliation state and degree of aggregation. Theoretical calculations predict a maximum surface area of approximately 2630 m² g⁻¹ for an ideal single-layer graphene sheet. However, experimentally measured values for GO are typically lower, ranging between 100 and 800 m² g⁻¹, owing to partial restacking and residual functionalization. Reduction processes can increase the surface area by removing oxygen groups and restoring planarity, although excessive reduction often promotes agglomeration, leading to a decrease in accessible surface area [56,58,59]. Brunauer–Emmett–Teller (BET) analysis is commonly employed to quantify this property, providing insights into pore size distribution and accessible surface sites. The surface area also correlates with adsorption capacity, electrochemical activity, and catalytic performance, making it a critical design parameter for functional applications.

GO is a chemically complex, 2D material derived from the oxidative exfoliation of graphite. Its chemistry is governed by a heterogeneous distribution of oxygen-containing functional groups and a partially disrupted sp² carbon lattice, which together define its physicochemical, electronic, and interfacial properties. The intricate balance between sp²-hybridized aromatic domains and sp³-hybridized oxidized regions makes GO a versatile precursor for graphene-based materials and a multifunctional platform for applications in energy storage, catalysis, and environmental remediation.

The chemical structure of GO consists of various oxygen functionalities distributed across the basal plane and sheet edges. Hydroxyl (–OH) and epoxy (C–O–C) groups are predominantly located on the basal plane, whereas carbonyl (C=O) and carboxyl (–COOH) groups are typically situated along the edges [60]. This spatial segregation results from the oxidative cleavage of C=C bonds within the graphite lattice during synthesis (e.g., Hummer’s or modified Hummer’s method). The coexistence of sp² and sp³ domains generates an inhomogeneous electronic landscape where the sp³ regions maintain delocalized π–conjugation, while sp³ zones act as electron-scattering centers, widening the band gap and modulating electrical conductivity [61]. The degree of oxidation in GO strongly influences its chemical reactivity and restoration potential toward rGO. Oxidation introduces oxygenated groups that enhance hydrophilicity and colloidal stability but disrupt electrical conductivity. Conversely, chemical, thermal, or electrochemical reduction processes remove oxygen species, partially recovering the conjugated carbon framework. The reduction mechanism often proceeds through dehydration, deoxygenation, and decarboxylation, accompanied by the restoration of sp² carbon domains [62,63,64]. Controlling this oxidation–reduction equilibrium enables the tuning of GO properties for targeted applications, from insulating films to semiconducting or conductive nanostructures. The acidic nature of GO arises primarily from carboxylic and phenolic groups, which can undergo protonation–deprotonation equilibria depending on the pH of the medium. This feature imparts GO with cation exchange capacity and metal ion coordination ability, making it suitable for adsorption and catalytic applications. Studies by Dimiev et al. [63] and others have shown that GO behaves as a dynamic acid, exhibiting reversible structural transformations under acidic and basic conditions, a behavior described by the DSM. The interplay between surface acidity and hydration further contributes to the material’s colloidal stability and reactivity. The abundance of oxygenated functionalities provides reactive sites for covalent functionalization, allowing GO to serve as a chemically tunable platform. Common modification routes include nucleophilic substitution, amidation, esterification, and silanization, which enable the attachment of polymers, metal nanoparticles, or biomolecules. Noncovalent interactions—such as π–π stacking, hydrogen bonding, and electrostatic attraction—also play crucial roles in forming hybrid composites. These chemical versatility pathways underpin GO adaptability in diverse fields including biomedicine, electronics, and environmental science.

The chemistry of GO can thus be summarized as a balance between oxidation-induced functional diversity and the preservation of graphitic order. Its amphiphilic nature, tunable oxygen content, and surface reactivity enable both bottom-up and top-down material design strategies. Understanding the underlying chemical principles governing GO’s formation, stability, and reactivity remains fundamental to advancing graphene-based technologies and optimizing material performance across various disciplines.

3.3. Synthesis Methods of GO

GO represents a chemically versatile derivative of graphite (Figure 5A), synthesized through various oxidation and exfoliation strategies developed over nearly two centuries. Its synthesis history reflects a progressive refinement of methods aimed at achieving higher oxidation efficiency, safety, reproducibility, and environmental sustainability. Fundamentally, GO is obtained through the oxidative disruption of the graphite lattice, introducing oxygen-containing functionalities such as hydroxyl, epoxy, carbonyl, and carboxyl groups, which facilitate layer separation and water dispersibility. Over time, numerous synthetic routes have been established, including the Brodie, Staudenmaier, and Hofmann methods (Figure 5B), as well as the more widely used Hummers and modified Hummers approaches [65,66,67,68,69,70,71]. More recent innovations include Tour’s method and electrochemical exfoliation processes, which seek to overcome the limitations of traditional oxidative pathways [72,73].

In 1859, Brodie conducted the first known synthesis of graphite oxide (GrO), later recognized as GO, by reacting graphite with nitric acid and potassium chlorate at 60 °C, over four days. He referred to the resulting compound as “graphitic acid.” However, the process presented significant explosion hazards due to the use of potassium chlorate and the slow recovery procedure. Despite these risks, Brodie’s method successfully achieved oxidation-induced delamination of rGO layers into graphene sheets. Building upon Brodie’s work, Staudenmaier sought to improve both the safety and efficiency of the synthesis. He introduced a mixed acid system, replacing pure nitric acid with a combination of sulfuric and nitric acids, thereby enhancing oxidation control and reducing the risk of detonation [75]. The Staudenmaier–Hofmann–Hamdi method (SHHM) later evolved from this approach, involving the treatment of graphite with potassium chlorate in a sulfuric–nitric acid mixture at approximately 90 °C for four days. This modification consolidated oxidation into a single step, increasing yield and uniformity. Nonetheless, the process remained hazardous due to the use of potassium chlorate and high temperatures. Numerous researchers—including Berthelot, Luzi, Charpy, Weinschenk, Kohlschütter, and Haenni—attempted to mitigate the explosive nature of the reaction through procedural adjustments, though without major success. To improve safety, concentrated sulfuric and nitric acids were retained as milder oxidizing agents, while potassium chlorate continued to serve as an in situ dioxygen source due to its strong oxidative power. Despite its limitations, this family of methods has remained a foundational route for preparing GO and rGO, and continues to be employed—with appropriate safety adaptations—by modern research groups [67].

The Hummer’s method (1958) marked a turning point in GO synthesis (Figure 5C), employing potassium permanganate (KMnO₄) and sodium nitrate (NaNO₃) in concentrated sulfuric acid (H₂SO₄) [75]. This approach eliminated the need for chlorate oxidants and reduced the formation of toxic gases such as ClO₂, significantly improving safety and efficiency. Despite these advantages, the conventional Hummers method produces large quantities of Mn²⁺-based waste and introduces nitrate contaminants, motivating further modifications. In response, modified Hummer’s methods have been developed to enhance oxidation efficiency, reduce environmental hazards, and control defect density [76]. Common improvements include eliminating NaNO₃, substituting KMnO₄ with less aggressive oxidants, or introducing phosphoric acid (H₃PO₄) as a co-solvent to yield GO with a more controlled oxidation level and fewer structural defects. The Tour approach (2010s) exemplifies this class, using a mixture of concentrated sulfuric and phosphoric acids with KMnO₄ as the oxidant, producing high-quality GO with fewer lattice defects and minimal toxic byproducts.

In parallel, electrochemical exfoliation and oxidation methods (Figure 5D) have emerged as greener and more controllable alternatives, avoiding the use of strong acids and heavy metal oxidants altogether. These approaches rely on anodic oxidation and intercalation of ions into graphite electrodes under applied voltage in aqueous electrolytes, enabling direct, scalable, and environmentally friendly production of GO or graphene-like materials [77].

Beyond the general overview of GO synthesis routes, it is important to briefly contrast the traditional Hummers method with its widely used modifications and with emerging greener or electrochemical approaches. The conventional Hummers route offers high oxidation efficiency but relies on strong oxidants (KMnO₄, NaNO₃, concentrated acids), raising safety and environmental concerns and limiting scalability. Modified Hummers protocols improve oxidation control and reduce toxic byproducts, yet still depend on hazardous reagents. In contrast, greener and electrochemical oxidation methods eliminate or minimize the use of strong oxidants, generate fewer secondary wastes, and offer more tunable oxidation levels, although their large-scale reproducibility and throughput remain under development. Together, these differences highlight the trade-offs between efficiency, environmental impact, and industrial feasibility across synthesis routes.

4. Graphene Oxide Membrane Technology

A membrane acts as a selective barrier that allows certain molecules to pass while blocking others, depending on the membrane’s structure and the properties of the species involved. Membrane technology has become a rapidly advancing field with critical applications in water purification, desalination, and chemical separations. Conventional membranes are generally classified as polymeric or inorganic. Polymeric membranes are favored for their high perm-selectivity, flexibility, and low cost; however, they suffer from limited resistance to heat, solvents, and corrosive environments. In contrast, inorganic membranes—such as ceramic, silica, and zeolite types—offer superior thermal and mechanical stability but often exhibit low permeability and complex synthesis requirements, especially when precise sub-nanometer pore control is needed. Although zeolite membranes present molecular-level sieving capabilities, their intrinsic catalytic and ion-exchange properties can interfere with neutral filtration processes.

Over the past decade, carbon-based nanomaterials like CNTs and, more recently, graphene and GO have opened new possibilities for membrane design. CNT membranes provided insight into nanoscale transport phenomena but faced fabrication challenges and scalability issues. The discovery of graphene in 2004 marked a turning point, revealing an atomically thin material with exceptional mechanical strength and chemical stability. However, pristine graphene is impermeable to gases and liquids due to its densely packed lattice. To overcome this limitation, GO with its oxygen-containing functional groups and tunable interlayer spacing has emerged as a highly promising membrane material. GO membranes demonstrate remarkable permeability to water while effectively rejecting ions and organic molecules, making them ideal candidates for next-generation separation technologies in both gaseous and liquid media.

4.1. Principles, Optimization Strategies and Transport Mechanisms of GO Membranes

GO membranes represent a versatile class of 2D nanomaterials with unique structural and chemical properties that enable selective molecular transport. Their performance is governed by the interplay between the membrane’s layered architecture, surface chemistry, and tunable transport pathways. GO membranes are typically constructed by stacking individual nanosheets into a laminate structure, forming interlayer nanochannels that serve as conduits for water, ions, and small molecules. These channels are defined by the spacing between adjacent sheets, which can be precisely adjusted through environmental conditions such as humidity and pH, or by chemical modifications including crosslinking and intercalation. The nanosheets themselves are functionalized with oxygen-containing functional groups—hydroxyl, epoxy, and carboxyl—that impart hydrophilicity and facilitate electrostatic interactions. These features collectively enable GO membranes to exhibit high permeability and selectivity [78,79,80].

Laminar GO membranes therefore feature two primary types of molecular transport channels: interlayer nanochannels formed between adjacent GO nanosheets, and intrinsic pathways arising from defects, pores, and sheet edges. Studies have shown that precise angstrom-level control over these transport routes—achieved through techniques such as chemical reduction, molecular or cationic cross-linking, intercalation, physical confinement, electric field modulation, pore engineering, and defect sealing—can significantly enhance membrane separation performance [81]. The microstructure of GO-based membranes can be broadly categorized into three types [82]:

(I)

Porous single-layer membranes, which consist of isolated GO sheets with engineered pores;

(II)

Few-layer laminated membranes, offering controlled spacing and high flux;

(III)

Multi-layer composite membranes, where GO is integrated with polymers or other nanomaterials to enhance mechanical and chemical stability.

Transport through GO membranes occurs via several mechanisms. Transport in GO membranes is governed by the principles of frictionless flow through sub-nanometer channels and electrostatic interactions and size-sieving effects, with the overall transport pathways being a combination of interlayer channels and intrinsic defects. Size exclusion plays a central role, as molecules larger than the interlayer spacing are effectively blocked. Charge-based selectivity is another key principle, with the negatively charged GO surface repelling anions and attracting cations, an effect that can be modulated by the surrounding ionic environment. Water transport is facilitated by hydration layers within the hydrophilic nanochannels, allowing rapid permeation. Additionally, GO membranes can conduct protons through hopping mechanisms, making them suitable for electrochemical applications such as fuel cells.

GO membranes exhibit a unique combination of transport mechanisms—size exclusion, electrostatic repulsion, and selective diffusion—enabled by their layered nanostructure (Figure 6). These membranes are composed of stacked GO nanosheets forming interlayer galleries that act as tunable nanochannels.

In GO membranes, nanosheets are restacked to form slit-like pores, with pore dimensions determined by the lateral size of the sheets and their interlayer spacing, known as d-spacing. This spacing, measured by X-ray diffraction, refers to the center-to-center distance between adjacent graphene planes. Pure graphene channels have a d-spacing of approximately 0.34 nm, while dry GO membranes typically exhibit a spacing around 0.8 nm. In hydrated conditions, the interlayer spacing typically ranges from 0.6 to 1.2 nm, allowing controlled permeation of water molecules while effectively rejecting ions, organic dyes, and gases. Membrane performance can be tuned through chemical reduction, crosslinking, or functionalization, which adjust both spacing and surface energy (Figure 7). The transport behavior is strongly influenced by the presence of oxygen-containing functional groups such as hydroxyl, carboxyl, and epoxy, which impart hydrophilicity and modulate the interlayer distance. Additionally, the performance of GO-based nanofiltration membranes is largely determined by their interlayer spacing, which is controlled by the interplay between GO swelling in the liquid phase and the adsorption of penetrant molecules [34]. In contrast, ions and larger organic molecules are hindered by steric constraints, electrostatic repulsion, and adsorptive interactions [83,84].

Structural modeling suggests that the spacing results from hydrogen bonding between functional groups, with contributions from the carbon–oxygen bond length (~0.22 nm) and the van der Waals radius of oxygen atoms (~0.17 nm), yielding a calculated d–spacing of ~0.78 nm closely matching experimental values. Unlike graphite, whose tight interlayer spacing prevents molecular permeation due to π–π stacking and electron cloud delocalization, GO membranes offer a free spacing of ~0.46 nm between sheets, forming hydrophobic nanocapillaries in pristine regions. This structural configuration is critical for understanding water and ion transport behavior and for developing accurate simulation models of GO membrane performance [84]. For gas separation, low-defect GO membranes with narrow interlayer gaps (<0.7 nm) are preferred, while water nanofiltration often uses membranes with expanded spacing (~1.0 nm) under solvent-swollen conditions. Small-scale fabrication methods offer better control and fewer defects, making them suitable for research. However, industrial adoption requires scalable techniques like bar-coating and spray coating, highlighting the gap between lab-scale precision and commercial production demands [85].

GO nanosheets are typically synthesized through strong oxidation of graphite under acidic conditions, resulting in a heterogeneous structure composed of both highly oxidized and non-oxidized regions. The oxygen-containing functional groups introduced during this process increase the spacing between adjacent nanosheets, forming interlayer galleries that serve as transport pathways in laminar GO membranes. These functional groups also enhance water permeability by promoting hydration. In contrast, the non-oxidized graphitic domains create capillary-like channels that enable nearly frictionless water flow. In their dry state, GO membranes exhibit an interlayer spacing of approximately 0.8 nm. Upon immersion in aqueous environments, this spacing can expand significantly, often exceeding 0.13 nm and, in some cases, reaching up to 6–7 nm due to the absorption of water molecules. While this swelling facilitates molecular and ionic transport, it also presents a challenge for salt rejection, as the expanded channels may exceed the size of hydrated ions such as sodium and potassium [86].

Water and ion transport through GO membranes occurs via three main pathways: through basal plane defects, interridge gaps between adjacent nanosheets, and the interlayer spacing between restacked sheets. While defect and interridge pores are difficult to control uniformly, the interlayer spacing—ranging from 0.34 to over 1 nm—can be precisely tuned. This spacing forms 2D graphene nanocapillaries that dominate the transport process, especially in thicker membranes. As such, GO membranes offer a valuable platform for investigating molecular transport through sub-2 nm channels. This structural configuration is critical for understanding water and ion transport behavior and for developing accurate simulation models of GO membrane performance. During permeation, molecules typically enter through defects, pores, or edges in the GO sheets before diffusing through the interlayer galleries (Figure 8).

Recent studies and reviews have emphasized the importance of understanding GO physicochemical properties, fabrication methods, and separation mechanisms. A key insight is that the ability to finely tune molecular transport pathways is fundamental to realizing the full potential of GO membranes in advanced separation applications [81]. This dual-domain structure—comprising hydrophilic oxidized regions and hydrophobic graphitic zones—creates complementary pathways that enhance both selectivity and permeability. Moreover, membrane performance can be fine-tuned through chemical reduction, crosslinking, or functionalization, which alter the interlayer spacing and surface energy, thereby optimizing separation efficiency for specific applications [87]. Optimization strategies focus on modifying the oxidation level of GO to adjust hydrophilicity and interlayer spacing, intercalating functional molecules to tailor transport behavior, and forming composites with other materials to enhance performance. These approaches allow GO membranes to be adapted for a wide range of applications, including water purification, gas separation, energy conversion, and biomedical devices.

GO membranes operate based on the synergistic effects of molecular sieving, electrostatic interactions, and capillary-driven transport through their unique layered nanostructure. The presence of oxygen-containing functional groups (hydroxyl, carboxyl, and epoxy) introduces hydrophilicity and tunable interlayer spacing, enabling selective permeation of molecules or ions while blocking larger species. Water molecules rapidly permeate through the 2D capillaries formed between stacked GO sheets due to hydrogen bonding with the oxygen groups and the low friction flow along the unoxidized graphitic regions. In contrast, ions and organic molecules are hindered by size exclusion, electrostatic repulsion, or adsorptive interactions. The balance between hydrophilic (oxidized) and hydrophobic (graphitic) domains provides a dual transport pathway that enhances both selectivity and permeability [88].

GO membranes have gained significant attention for gas separation due to their sub-nanometer channels that enable molecular sieving. Research has explored how interlayer spacing, structural defects, and flake size influence gas permeation. Notably, smaller GO flakes tend to reduce gas permeance, suggesting dual transport pathways through defects and nanochannels [89]. Quantifying gas transport requires understanding both diffusivity and solubility within multilayer structures. However, experimental data remain limited and inconsistent, especially for thick membranes (>10 μm). Molecular dynamics simulations offer higher diffusivity estimates, comparable to liquid-phase diffusion. Still, the lack of standardized characterization and the complexity of defect-driven models pose challenges for practical application of GO membranes in gas separation. Thus, GO membranes act as dynamic nanofluidic channels, where molecular transport is governed by both physical confinement and chemical interactions, making them highly adaptable for water purification, gas separation, and ion sieving applications.

4.2. Fabrication Techniques

Fabrication of GO membranes typically begins with stable aqueous dispersions of GO nanosheets. These dispersions are processed using techniques such as vacuum filtration, spin coating, drop casting, or layer-by-layer assembly to form thin films with controlled thickness and uniformity. The membranes are often supported on porous substrates—such as polymer filters or anodic aluminum oxide—to improve mechanical strength and operational stability. Chemical treatments, including crosslinking with divalent cations or reduction of oxygen groups, are employed to fine-tune membrane properties such as selectivity, conductivity, and durability [90].

Although GO films are utilized in various applications, such as electrodes, common distinction focuses exclusively on their role in separation technologies. The aim is to evaluate the advantages and limitations of different fabrication methods to inform the efficient synthesis of GO membranes. Of the eight common techniques, five—vacuum filtration, spin coating, pressure-assisted assembly, dip coating, and drop-casting—are categorized as small-scale, batch-based approaches. In contrast, spray coating, bar/doctor blade coating, and slot-die coating are considered large-scale methods capable of continuous membrane production. In these cases, the fabrication area is constrained only by the physical dimensions of the equipment.

4.2.1. Vacuum Filtration/Pressure-Assisted Assembly

Vacuum filtration is the most widely used method for fabricating GO membranes, particularly for separation applications. It involves filtering a diluted GO dispersion through a porous substrate under vacuum, which deposits GO flakes to form a selective layer. Membrane thickness can be tuned from a few nanometers to several micrometers by adjusting the dispersion concentration. Thicker films can be delaminated to create freestanding structures, though they often exhibit disordered morphology due to uneven pressure distribution and require longer filtration times, making large-scale production challenging [91]. Despite its simplicity, vacuum filtration consumes significant solvent volumes and yields membranes limited to centimeter-scale dimensions. Thin GO membranes (~10 nm) have demonstrated high organic solvent permeance and effective dye rejection due to edge pinholes and minimal thickness. Solvation of rGO flakes before drying can expand interlayer spacing and enhance permeance, though this approach is mainly suitable for lab-scale fabrication [92]. The method also enables hybrid membrane designs. For instance, mixing rGO with GO nanoribbons (GONRs) increases water flux by creating enlarged nanochannels [93]. Substrate morphology plays a crucial role: wrinkled surfaces can disrupt GO stacking and widen interlayer spacing, improving water transport. Additionally, nanoporous rGO sheets prepared via thermal treatment and functionalization have shown exceptional water permeance (up to 586 LMH/bar) [94]. Solvent choice is critical, especially for graphene materials with low oxygen content. Organic solvents like dimethylformamide and N–methyl–2–pyrrolidone can dissolve polymer supports, necessitating either functionalization for water solubility or the use of eco-friendly alternatives. Pressure-assisted assembly is a membrane fabrication technique that, like vacuum filtration, relies on a filtration process to deposit GO layers. However, instead of applying a vacuum beneath the substrate, this method uses positive pressure directly on the GO dispersion. The applied pressure drives the dispersion through the porous support, enabling the formation of a selective carbon layer on the surface. Vacuum filtration produces dense, layered films with uniform interlayer spacing and high rejection efficiency.

4.2.2. Spin-Coating/Dip Coating/Drop-Casting Enables Precise Control of Thickness for Ultrathin and Flexible Membranes

Spin-coating, dip coating, and drop-casting are three solution-based techniques commonly used for fabricating GO membranes, each with distinct characteristics. Spin-coating enables the formation of thin, well-aligned GO layers with high selectivity and uniform morphology, making it suitable for forward osmosis and ion rectification applications. However, it is limited by substrate size and requires smooth surfaces to avoid defects [95]. Dip coating offers greater flexibility in substrate geometry and is relatively insensitive to shape, relying on immersion and withdrawal from GO dispersions. While achieving uniform alignment can be challenging, optimization of parameters such as temperature, concentration, and withdrawal speed has led to successful implementations in nanofiltration and oil–water separation [96]. Drop-casting is the most rudimentary method, involving the deposition of GO dispersions directly onto substrates followed by drying. Due to the absence of surface ordering forces, it is difficult to achieve well-aligned structures, often resulting in coarse membrane surfaces and limited quality control. Nevertheless, drop casting has been used to fabricate freestanding membranes for ion sieving and pervaporation, with studies demonstrating selective separation of heavy metals and organic contaminants. Despite its simplicity, drop casting faces challenges in scaling up membrane dimensions and ensuring reproducibility, as the process is highly dependent on solvent evaporation dynamics. Together, these methods form a versatile toolkit for GO membrane fabrication, balancing precision, adaptability, and scalability across different application contexts [62].

4.2.3. Spray Coating, Bar/Doctor Blade Coating, and Slot-Die Coating as Scalable Methods for GO Membrane Fabrication

Spray coating, bar/doctor blade coating, and slot-die coating are three scalable techniques widely employed in the fabrication of GO membranes. Spray coating disperses GO droplets onto substrates using a spray gun, forming membranes as the solvent evaporates. It is relatively insensitive to substrate shape and suitable for complex geometries, though membrane uniformity may suffer. Studies have shown that spray-coated membranes exhibit reduced wrinkle formation, which narrows interlayer spacing and enhances selectivity [97]. Layer-by-layer deposition using oppositely charged GO variants improves adhesion and barrier properties, and successful applications include hollow fiber nanofiltration membranes with high flux and dye rejection. Bar and doctor blade coating utilize concentrated GO dispersions with viscoelastic properties, enabling shear-induced alignment of GO laminates [98]. These methods produce high-quality membranes with thicknesses down to 100 nm, though ultrathin layers are more difficult to achieve. They require flat substrates and have been enhanced through gravure printing, ionic liquid additives, and nanoporous GO structures, yielding membranes with strong performance in nanofiltration and organic solvent separation. Slot–die coating, meanwhile, offers continuous membrane production with precise control over layer thickness and uniformity. It is particularly suited for large-area fabrication and roll-to-roll processing, making it ideal for industrial-scale applications. Like bar coating, it benefits from concentrated GO dispersions and shear alignment, but with greater automation and reproducibility. Together, these scalable methods provide a robust platform for translating GO membrane technologies from lab-scale innovation to commercial deployment [99].

4.2.4. Layer-by-Layer Assembly (LbL) Allows Nanometer-Scale Tuning of Multilayered Architectures and Integration with Polymers

Layer-by-layer (LbL) assembly is a versatile and tunable method for fabricating GO membranes, offering precise control over thickness, composition, and interfacial properties. This technique involves the sequential deposition of oppositely charged materials—typically pristine GO and functionalized GO variants—onto a substrate, forming highly ordered multilayer structures. The electrostatic interactions between layers enhance adhesion and structural integrity, enabling the creation of membranes with tailored separation performance [100]. Heo et al. demonstrated an LbL spray-coating approach using negatively charged GO and positively charged amine-functionalized GO to fabricate oxygen barrier films with improved adhesion and reduced gas permeability [101]. Gu et al. applied LbL assembly to produce chemically resistant GO membranes for dye separation and desalination, achieving long-term stability and high rejection rates. GO membranes were constructed on modified Nylon–6 substrates using alternating layers of positively charged guanidine-functionalized GO (P–GO) and negatively charged sulfonated GO (S–GO). This approach allows precise control over membrane architecture by tuning the number of layers and surface charge to target specific contaminants. The resulting ML11 membrane exhibited exceptional dye rejection (>98%) and strong desalination performance, with 93.4% rejection of Na₂SO₄ and 72.7% for NaCl. These membranes maintained consistent separation efficiency for over 168 h under cross-flow conditions and demonstrated robust resistance to chemical corrosion and bacterial degradation, making them suitable for long-term operation in harsh environments [102]. The LbL method is compatible with various substrates and scalable formats, including hollow fibers and flat sheets, and can be integrated with interfacial polymerization to form composite nanofiltration membranes. While LbL fabrication offers excellent control and performance, it requires careful optimization of deposition cycles, pH, and ionic strength to ensure uniformity and reproducibility. Overall, LbL assembly stands out as a promising route for engineering high-performance GO membranes for water purification and molecular separation [103].

4.2.5. Interfacial Self-Assembly: Forms Defect-Free, Large-Area GO Films at Liquid Interfaces

Interfacial self-assembly is a method that uses liquid interfaces to create large-area, defect-free films of GO. This is achieved by confining GO particles to the interface between two immiscible liquids, where they can self-assemble into highly ordered, 2D structures driven by the minimization of free energy. The resulting films can be transferred to a substrate, and the process allows for the creation of membranes with tunable properties for applications like water purification [104].

4.2.6. Mixed Matrix Membranes (MMMs) Combine GO Nanosheets with Polymer Matrices for Enhanced Mechanical Strength and Permeability

MMMs combine polymer matrices with inorganic or hybrid fillers to enhance separation performance, mechanical strength, and chemical stability. The review highlights how filler selection—such as zeolites, MOFs, carbon-based nanomaterials, and GO—can tailor membrane properties for specific applications like gas separation, pervaporation, and water treatment. Key challenges include achieving uniform filler dispersion, strong interfacial compatibility, and scalable fabrication [105]. GO MMMs are fabricated by incorporating GO nanoparticles into a polymer matrix, typically polyethersulfone (PES), using phase inversion techniques. This approach enhances membrane hydrophilicity, mechanical strength, and antifouling properties. Optimal GO loading (0.5 wt%) significantly improves water flux and dye rejection, achieving over 99% removal for dyes like Acid Black and Rose Bengal. These membranes also demonstrate long-term operational stability and resistance to chemical and biological degradation, making them highly effective for industrial wastewater treatment applications [106]. A novel polyamide@GO MMMs is designed for efficient molecular sieving of nitrogen (N₂) from volatile organic compounds (VOCs), such as cyclohexane. Unlike conventional MMMs formed by mechanical mixing, this membrane was synthesized via in situ polymerization, chemically linking GO with polyamide using 2,6,14–triaminotriptycene and octanedioyl chloride. This method improved filler dispersion and created more amorphous regions, enhancing gas transport channels [107]. Assembly of GO at the air–liquid interface has emerged as an effective strategy to produce well-ordered nanoflakes for membrane modification. For example, Bonatout et al. demonstrated that GO spontaneously forms bilayer structures at the air–water interface, enabling precise control over sheet density during transfer via Langmuir–Blodgett or Schaefer techniques [108]. Petukhov et al. further showed that such aminated GO films can be transferred onto polyamide membranes to enhance hydrophilicity, reduce surface roughness, and improve antifouling performance, with high flux recovery and scalable processing [109].

4.3. Types of GO Membranes

Depending on the structural design and the presence or absence of a supporting substrate, GO membranes can generally be classified into three main types: free-standing GO membranes, supported-GO membranes, and GO-modified composite membranes (Figure 9). Each type exhibits distinctive structural characteristics, preparation methods, and performance behaviors, which determine their suitability for various separation and purification applications. Free-standing GO membranes utilize stacked GO nanosheets directly as the separation layer. In contrast, supported-GO membranes consist of a GO layer deposited onto a polymeric or inorganic substrate, where the GO serves as the active separation interface. GO-modified composite membranes are fabricated either by blending GO nanosheets into polymer casting solutions or by post-coating pre-formed membranes with GO to enhance surface functionality [110].

4.3.1. Free-Standing GO Membranes

Free-standing GO membranes are composed solely of stacked GO nanosheets without any supporting material. They are typically prepared by vacuum filtration, drop-casting, or layer-by-layer deposition of GO dispersions, followed by drying to form flexible films. Due to their layered structure and well-controlled interlayer spacing, these membranes demonstrate exceptional molecular sieving and selective permeability. However, they may suffer from mechanical fragility, swelling in aqueous environments, and limited scalability [111].

4.3.2. Supported-GO Membranes

Supported-GO membranes are fabricated by depositing GO layers onto porous substrates such as PES, polyvinylidene fluoride (PVDF), ceramic, or metal oxide supports. The underlying substrate provides mechanical reinforcement, improves handling stability, and prevents membrane rupture under pressure-driven operations. GO is typically coated via spin coating, vacuum filtration, spray coating, or dip coating techniques. These membranes offer a good balance between mechanical robustness and separation performance. By depositing GO layers onto polymeric or inorganic substrates, researchers have significantly enhanced both separation performance and structural stability. For instance, Hung et al. prepared GO-modified PAN membranes via PASA, achieving high selectivity for IPA/water pervaporation due to densely packed GO laminates [112]. Chu et al. coated PES membranes with GO, improving humic acid rejection and water flux through strong hydrogen bonding between GO and PES [113]. Shin at al. report a scalable and stable casting of GO dispersion onto PES membranes for water–ethanol separation, molecular-level insights into selective transport, and the feasibility of GO/PES membranes for dehydration of biofuels [114].

4.3.3. GO-Modified Composite Membranes

GO-modified composite membranes are typically polymer-based membranes incorporated or surface-modified with GO nanosheets to enhance hydrophilicity, antifouling behavior, and selectivity. GO acts as a functional additive or coating, improving water permeability, chemical resistance, and selective transport properties through electrostatic and hydrogen–bonding interactions. Such membranes can be prepared by phase inversion, interfacial polymerization, or blending methods, allowing scalable production and tunable performance. A wide range of GO–polymer composite membranes has now been developed, demonstrating enhanced water permeability, selectivity, and antimicrobial properties [115,116]. These modifications generally follow two main strategies. In the first, GO nanosheets are blended into polymer casting solutions during membrane fabrication, allowing for uniform dispersion throughout the membrane matrix. In the second, GO nanosheets are applied onto the surface of preformed polymeric membranes, enabling functionalization via surface modification techniques [117].

4.4. Materials Suitable for GO Membranes

The materials used for GO membranes vary depending on the membrane structure, whether free-standing, supported, or composite, and the intended application, such as water purification, gas separation, or pervaporation (Figure 10). At the core of all GO membranes is graphene oxide itself, composed of oxidized graphene sheets rich in oxygen-containing functional groups, which confer hydrophilicity, tunable interlayer spacing, and chemical modifiability for enhanced selectivity [81,118]. To improve mechanical strength and flexibility, polymer supports are commonly used, including PES for its thermal and chemical stability, polysulfone (PSf) for its processability, polyacrylonitrile (PAN) for its compatible surface chemistry, PVDF as a hydrophobic base, and cellulose acetate (CA) for its eco-friendly and hydrophilic nature [113,115]. Inorganic supports such as alumina (Al₂O₃), titania (TiO₂), silica (SiO₂), and porous ceramics offer structural rigidity and high chemical resistance [119]. GO membranes can also be enhanced through hybridization with nanomaterials like rGO for improved conductivity, MOFs for increased porosity and selectivity, CNTs for mechanical reinforcement, and layered double hydroxides (LDHs) for ion exchange and stability [92,117]. Finally, crosslinkers and modifiers such as ethylenediamine (EDA), glutaraldehyde, polyethyleneimine (PEI), multivalent metal ions (e.g., Fe³⁺, Ca²⁺, Al³⁺), and various polymers or surfactants are employed to stabilize GO layers, tune interlayer spacing, and enhance adhesion and dispersibility [120].

5. Application of GO in Membrane Technology

Freshwater is one of the most essential resources sustaining all forms of life on Earth. However, despite the planet’s abundant water reserves, less than 3% of the total water is freshwater, and an even smaller fraction is readily accessible for human use. The rapid growth of populations, industrialization, and agricultural activities has led to the depletion and contamination of natural water sources, creating a global challenge in providing clean and safe water. Water contaminants generally fall into several categories, including suspended solids, organic compounds, heavy metals, microorganisms, and dissolved salts. These pollutants not only degrade water quality but also pose serious environmental and health risks. To address these issues, various water purification methods have been developed, such as filtration, adsorption, chemical coagulation, photocatalysis, and membrane-based separation. Among them, membrane technologies have gained increasing attention due to their efficiency, scalability, and ability to achieve high levels of contaminant removal without requiring additional chemical reagents. Conventional membranes, however, often suffer from limitations such as low permeability, fouling, and insufficient selectivity, which hinder their large-scale application in water treatment and desalination [121,122].

In recent years, the emergence of GO has opened new possibilities for advanced membrane technologies. Owing to its unique 2D structure, tunable surface chemistry, mechanical robustness, and abundant oxygen-containing functional groups, GO exhibits exceptional potential as a membrane material for both water purification and desalination (Figure 11A). Therefore, GO-based membranes represent a significant advancement toward next-generation filtration and desalination technologies, combining high selectivity, permeability, and mechanical strength with the potential for large-scale, cost-effective production [123,124].

Individual GO sheets are composed of hydrophobic sp² carbon regions and functionalized sp³ domains containing oxygen-bearing groups. Typically, water transport occurs more readily through the hydrophobic sp² regions of the basal plane, while the presence of functional groups in the sp³ domains introduces hydrogen bonding and electrostatic interactions that can hinder molecular mobility. However, these oxygenated groups also facilitate swelling and maintain interlayer spacing, thereby improving overall water transport. Consequently, the type and density of these functional groups critically influence membrane flux [123,126]. To disclose how diverse oxygen-containing functional groups affect GO nanomembranes on water purification performance, Yu et al. fabricated carboxyl-dominant GO (C–GO), hydroxyl-dominant GO (H–GO) and epoxy-dominant GO (E–GO), on PVDF support by the vacuum filtration method. The permeability of water of C–GO, H–GO and E–GO membranes was 9.19, 7.22 and 5.94 L m⁻² h⁻¹ bar⁻¹, correspondingly. Three different dyes (MB, EBT and TB) were used for estimation of the GO membranes’ purification performance. Authors demonstrated that GO membranes enriched with different dominant groups (–COOH, –OH, and –COC) exhibited water transport rates following the order COOH > OH > COC, consistent with their corresponding interlayer spacing. The enhanced permeability of COOH-rich GO membranes was further linked to the higher degree of oxidation, which induces additional structural defects that promote water flow [127]. Qui et al. applied molecular dynamics simulations to examine the effects of three common edge functional groups on graphene oxide (GO) nanosheets, carboxyl (–COOH), hydroxyl (–OH), and hydrogen (–H) [128]. A controlled numerical approach was used to isolate and evaluate the individual influence of each functional group. The findings revealed that the –COOH groups negatively affect water transport because their bulky geometry creates steric hindrance that impedes flow. In contrast, –OH groups enhance water transport by “pulling” water molecules through the nanosheet layer via strong hydrogen bonding interactions, while –H atoms also facilitate water movement due to their minimal steric resistance. Additionally, the size of the inter-edge channel significantly influences how these functional groups affect transport behavior. Overall, the results suggest that for designing GO membranes with high water flux, it is advantageous to minimize –COOH groups while maintaining a balanced presence of –OH and –H groups at the edges. Sun et al. investigated selective ion penetration of GO membranes (Figure 11B(a)) [125]. For the GO flakes’ preparation, they used modified Hummer’s method obtaining a GO suspension (Figure 11B(b)), where the GO sheets’ lateral size of several hundred nanometers was evaluated by AFM microscopy (Figure 11B(c)). Authors prepared GO membranes as free-standing films (Figure 11B(d)), where SEM analyses of the surface morphology and cross-sectional structure of the membranes (Figure 11B(e–g)) revealed that the fabricated GO membranes exhibit a distinctly wrinkled surface and a well-defined lamellar architecture. As shown in Figure 11B(f), the membrane thickness was below 10 μm. For the experiment of selective ion penetration (Figure 11C), authors used a plastic separator, as presented in Figure 11A. The conductivity measurements of 0.1 mol/L metallic salt solutions ranged from 6 to 29 mS/cm, while the filtrates showed values in the μS/cm range, indicating effective ion rejection by the GO membrane. As shown in Figure 11C(a), the conductivities of all salt solutions followed a similar trend, remaining stable initially and then increasing nearly linearly. However, permeability varied greatly among different salts: sodium salts permeated efficiently, whereas heavy-metal salts showed much slower diffusion. Strong acids and bases exhibited higher permeability than neutral salts, while basic salts such as NaHCO₃ showed lower permeability due to gas formation from reactions with GO’s functional groups. Notably, copper sulfate was completely blocked, with filtrate conductivity remaining identical to that of deionized water. The initial stage of conductivity variation (Figure 11C(b)) showed minimal change for most filtrates except NaHSO4, while a distinct peak appeared for NaOH due to its reaction with GO surface groups. Relative conductivities revealed that sodium salts permeated faster than heavy-metal salts, with rates following the order NaOH > NaHSO₄ > NaCl > NaHCO₃. For a 0.1 mg/mL rhodamine B solution, UV–vis spectra (Figure 11C(c)) showed no detectable dye in the filtrate, and the absorption spectra before and after 3 h overlapped, confirming that organic contaminants could not penetrate the GO membrane. The relative concentration plot (Figure 11C(d)) further confirmed a constant rhodamine B concentration, and the unchanged solution color (inset) highlighted the strong rejection capability of GO membranes. Finally, the permeability of NaCl and NaHSO₄ at varying concentrations (Figure 11C(e,f)) showed that reduced feed concentration led to decreased ion permeation. Below a critical threshold, the salts could no longer pass through the GO film. Based on the results obtained, the authors proposed a theoretical model for the selective permeation characteristics of GO membranes (Figure 11D) [125]. The oxygen-containing functional groups on GO sheets tend to aggregate, leaving behind non-oxidized graphitic regions that form a continuous two-dimensional network of graphene nanocapillaries (Figure 11D). These nanocapillaries generate high capillary pressure, enabling rapid, low-friction water transport through the hydrophobic channels. In contrast, water molecules within the oxidized regions move more slowly due to hydrogen bonding and interactions with surface functional groups. When the interlayer spacing between GO sheets increases, the nanocapillaries can accommodate hydrated ions, allowing partial ion permeation (Figure 11C). However, the ion selectivity of GO membranes cannot be explained solely by ionic radius. For instance, the observed ion permeation sequence, Na⁺ > Mn²⁺ > Cd²⁺ > Cu²⁺, does not correlate with the order of ionic or hydrated radii (Mn²⁺ > Cd²⁺ > Cu²⁺ > Na⁺). This discrepancy arises from the strong electrostatic and chemical interactions between the oxygenated functional groups on GO surfaces and the hydrated ions. Consequently, the selective ion transport through GO membranes results from the combined effects of nanocapillary confinement and ion–surface interactions rather than simple size exclusion [125,129,130].

Huang et al. presented a design of nanostrand-channelled graphene oxide (NSC–GO) ultrafiltration membranes, which feature a dense network of nanochannels approximately 3–5 nm in diameter. These membranes demonstrated exceptional small-molecule separation performance and ultrafast water permeation of ~695 ± 20 L m⁻² h⁻¹ bar⁻¹, which was nearly ten times higher than that of conventional GO membranes while maintaining excellent rejection efficiency. Remarkably, the water flux of NSC–GO membranes was also over two orders of magnitude greater than that of standard commercial ultrafiltration membranes with comparable selectivity. Interestingly, the separation behavior of these membranes displayed an unusual dependence on applied pressure, deviating from the response typically observed in porous polymeric membranes and indicating pressure-tunable (mechano-tunable) transport properties. Combined experimental analysis and molecular dynamics (MD) simulations suggested that this phenomenon resulted from the reversible deformation of the GO nanochannels; their geometry and cross–sectional area dynamically adjusted under varying pressure, influencing both water permeance and solute rejection. Moreover, water transport through NSC–GO membranes followed a classical viscous flow regime with significantly reduced boundary slip compared to the nearly plug-like flow observed in hydrophobic graphene-based channels such as single-walled carbon nanotubes and graphene membranes [131].

Yasmeen et al. employed GO nanomaterials for water purification, leveraging their exceptional mechanical strength and unique 2D structure [121]. GO nanoparticles were synthesized via Hummer’s method and incorporated into cellulose filter paper using a dip-coating technique, resulting in a GO-coated cellulose membrane with enhanced filtration capability. This newly developed membrane showed significant improvements in water quality across samples collected from urban, rural, and industrial regions of Sialkot. Key outcomes included a notable decrease in electrical conductivity (EC) and an improved Sodium Adsorption Ratio (SAR), both indicating superior filtration performance. The membrane effectively reduced sodium ion concentrations and mitigated salinity, making the treated water more suitable for agricultural applications. Changes in pH levels, especially pronounced in industrial areas, highlighted the impact of industrial waste on local water quality. Overall, these findings underscored the importance of continuous water quality monitoring and treatment, particularly in regions with diverse water sources. The integration of GO-based nanomembrane filtration demonstrates clearly the potential for enhancing water quality by lowering EC and improving SAR values, rendering water more appropriate for consumption and irrigation. Interestingly, pH values remained relatively stable before and after filtration, suggesting that while ionic and salinity parameters improved, acidity was largely unaffected. The research also revealed significant regional variations in pH, likely influenced by geological conditions, industrial emissions, and water treatment practices. A general decrease in pH across all sites indicated the possible introduction of acidic compounds, which may have environmental and health implications. Industrial zones, in particular, showed higher initial pH values followed by a decrease post-filtration, whereas elevated sodium and EC levels were most evident in urban and industrial regions, likely due to industrial discharges, agricultural runoff, or geological factors. Furthermore, GO demonstrated potential as a catalyst support for pollutant reduction, membrane separations, and chemical production, highlighting its versatility beyond filtration. While the study provides valuable insights into GO-based water treatment, further research is needed to identify specific pollution sources and evaluate the long-term environmental effects of these contaminants in Sialkot’s water systems [121].

Beyond the well-known permeability–selectivity trade-off, one of the major challenges facing GO membranes, both in aqueous and solvent-based environments, is their long-term structural stability. GO laminates, rich in oxygen-containing functional groups, typically exhibit an initial interlayer spacing of about 0.8 nm; however, this spacing is highly sensitive to the surrounding medium. When exposed to solvents, and particularly to water, these layers tend to swell significantly. Over extended hydration periods, dissociation of functional groups can occur, generating electrostatic repulsion that pushes the sheets apart, expanding the interlayer distance to as much as 6–7 nm. Such expansion not only undermines the size-sieving capability of the membrane but can also lead to mechanical destabilization and delamination of the GO stack itself. Mitigation strategies often involve the partial reduction of GO nanosheets, which improves structural rigidity and decreases swelling. However, this approach inevitably reduces hydrophilicity and water permeability, creating yet another delicate balance between performance and stability. Furthermore, GO membranes are commonly supported on porous substrates to enhance mechanical strength. Yet, for solvent-based separations, chemical compatibility between the support and the selective GO layer becomes critical. Many conventional polymeric supports, such as polysulfone, are unsuitable because they degrade or deform upon solvent exposure. As a result, researchers are turning toward chemically resistant materials, including ceramic supports and crosslinked polymeric substrates, which, though more expensive, provide the durability and chemical resilience needed for prolonged operation [132].

Thermoplastic polyurethane (TPU) is a widely used elastomeric polymer known for its versatility and tunable properties, which can be tailored by adjusting its monomer composition to meet specific application needs. However, conventional TPU often suffers from limited stiffness and mechanical strength, primarily due to its low hard-segment content, restricting its performance in demanding environments. To overcome these limitations, researchers have explored the incorporation of functional fillers, such as CNTs and graphene, into the TPU matrix to improve its mechanical, thermal, and electrical properties. Among these, GO-reinforced TPU composites have attracted particular attention because GO not only offers excellent mechanical reinforcement but also introduces functional groups that enhance interfacial compatibility and dispersion within the polymer matrix. This synergy results in composites with a unique balance of flexibility, strength, and multifunctionality, making GO–TPU systems promising candidates for advanced structural, sensing, and flexible electronic applications [133,134,135]. In their study, LV et al. demonstrated a graphene oxide-reinforced thermoplastic GO/TPU liquid-gating elastomeric porous membrane (LGEPM) with excellent stretchability, durability, and antifouling performance [136]. The membrane design integrated GO into the TPU matrix, where hydrogen bonding between the two components enhances mechanical strength and stability. Figure 12A shows the design strategy and working mechanism of the GO/TPU LGEPM. The GO/TPU elastomeric membrane (EM) was synthesized via a solution-blending method, forming hydrogen bonds between the oxygen groups of GO and the N–H groups of TPU, which markedly improve the composite’s tensile strength. Porous structures were then precisely fabricated using laser cutting to obtain the GO/TPU elastomeric porous membrane (EPM). The EPM was subsequently infused with a functional gating liquid to create the GO/TPU LGEPM, enabling a stable, high-performance antifouling liquid gating system. Compared with the pure TPU LGEPM, the GO/TPU LGEPM exhibits superior tensile strength and reversible stretchability. The infused liquid adapts to pressure changes, dynamically regulating gas/liquid transport. To ensure interfacial stability, the membrane and gating liquid were energy-matched, with silicone oil selected as the gating liquid for demonstration. Under mechanical stress, the GO/TPU LGEPM stretches and its pores expand, altering the critical pressures required for gas or liquid transport. Thus, by stretching and releasing the membrane, these pressures can be dynamically tuned (Figure 12A). Unlike conventional membranes that allow passive gas flow, liquid-gated membranes require both gases and liquids to overcome specific critical pressures to pass through. This enables effective gas–liquid separation. When the applied pressure (P) is below both critical pressures, no transport occurs (P < ΔP₍gas₎ < ΔP₍liquid₎). Upon stretching, the enlarged pores reduce these pressures so that only gas passes (ΔP₍gas₎ < P < ΔP₍liquid₎). With further stretching, both gas and liquid permeate (ΔP₍gas₎ < ΔP₍liquid₎ < P). Releasing the tension restores the membrane to its original state.

Lv et al. showed notable differences in the critical pressures of gas and liquid transport through the GO/TPU EPM with and without a gating liquid. The bare GO/TPU EPM exhibited a transmembrane pressure of 2318 ± 218 Pa for liquid transport, while the GO/TPU LGEPM reduced this value by 32% to 1560 ± 18 Pa, highlighting its energy-efficient performance. As shown in Figure 12B(a,b), the liquid gating mechanism enables tunable critical pressures for both gas and liquid transport. Since pore size strongly influences these pressures, the elastic GO/TPU matrix allows dynamic pore deformation, while the gating liquid adapts its shape to the applied stress. Consequently, stretching enlarges the pores and decreases the critical pressures of both gases and liquids. Simulations confirmed that at 100% strain, ΔP₍gas₎ decreases from 939 ± 48 Pa to 557 ± 22 Pa, and ΔP₍liquid₎ from 1366 ± 44 Pa to 799 ± 7 Pa (Figure 12B(c)). Upon release, the membrane fully recovers. Repeated cycling (10 cycles, 50% strain) demonstrated stable and reversible pressure responses, validating the durability of the GO/TPU LGEPM (Figure 12B(d)). Elzubair et al., in their study, fabricated GO membranes by depositing GO dispersions of varying concentrations onto commercially available porous polymeric supports, such as polytetrafluoroethylene (PTFE), using vacuum-assisted filtration [138]. The desalination performance was evaluated through filtration tests of NaCl solutions, with permeate conductivity used to assess salt rejection efficiency. As authors summarized, the GO membrane with a concentration of 0.6 mg/mL exhibited the highest water flux and permeability, along with the lowest conductivity and NaCl concentration, achieving superior salt rejection. Permeate conductivity closely matched that of deionized water, confirming effective ion exclusion. In contrast, the membrane prepared with 0.8 mg/mL GO showed reduced permeability, higher conductivity, and poorer rejection performance, likely due to its thicker structure. These results demonstrate that GO concentration, and thus membrane thickness, strongly influences both water permeability and NaCl rejection.

Precise control of interfacial interactions through surface chemistry engineering is essential for developing high-performance antifouling coatings and separation membranes. In their work, Yang et al. proposed a hydrophobic chain engineering strategy to tune membrane surfaces at the molecular level [137]. Hydrophilic phytic acid and hydrophobic perfluoro carboxylic acids were sequentially assembled on a GO membrane, creating an amphiphilic surface (Figure 12C). The perfluoroalkyl chains lowered surface energy, while varying their length adjusted surface hydration, synergistically enhancing both fouling resistance and fouling release. Figure 12D(a–d) show the water density profiles along the direction perpendicular to the surfaces. Owing to strong Coulombic interactions between the super hydrophilic substrate and water molecules, a dense interfacial water layer (~0.6 nm thick) forms on all surfaces. The sharpness of this density peak first decreases and then increases with the perfluoroalkyl chain length, indicating that the hydration layer structure initially loosens before becoming more ordered. Notably, the F₆–hGO surface exhibits a broader and more uniform hydration layer. Authors stated that previous studies have shown that moderate disruption of the hydration layer can enhance interfacial water content and thus strengthen hydration capacity. Surface hydration showed a nonlinear dependence on chain length (C₄–C₁₀), peaking at C₆ due to more uniform water orientation, as confirmed by molecular dynamics simulations. The resulting membrane achieved excellent antifouling performance (flux decline < 10%, flux recovery ~100%) and high permeance (~620 L m⁻² h⁻¹ bar⁻¹) in oil–water separation [137].

Table 1. GO membranes for water treatment.

Fabrication Process	Performance	Composition/Application	Ref
Vacuum filtration	312.8 LMH/bar	GONR mixture on a porous support for dye and ion separation	[93]
Spin-Coating	3.5 × 10−7 mol/(m2·s·Pa)	GO flakes on porous alumina substrate for gas separation	[139]
Dip Coating	56.8 to 330 LMH/MPa	polyamide-imide hollow fiber/GO membranes for Eriochrome black T	[140]
Spray coating	40 and 24 LMH/bar	GO layer on polyetherimide for Rose Bengal	[141]
Bar/Doctor Blade Coating	8 LMH/bar	GONR gels for dye molecules	[142]
Slot-Die Coating	30 LMH/bar	deoxygenated GO sheets for dye molecules	[99]
Layer-by-Layer Assembly	CO2/N2 selectivity of 48.48 N2 permeance at 1204.25 GPU	Tertiary-amine-stabilized gold nanoparticles and GO for CO2 capture	[101]
Interfacial Self-Assembly	191 L m−2 h−1 bar−1	GR/GO@PEI composite for Congo red	[143]
Mixed matrix membranes (MMMs)	2.94 L m−2 h−1	GO polyethersulfone (PES) for Hemodialysis	[144]

summarizes the correlation, from the selected literature papers, of the type of GO membrane with the methods of preparation, and application performance.

GO has emerged as a highly effective additive and structural component in proton and ion-exchange membranes due to its abundant oxygen-containing functional groups and tunable interlayer architecture. The presence of hydroxyl, epoxy, and carboxyl groups facilitates proton hopping via Grotthuss-type mechanisms, while the hydrophilic nanochannels formed between stacked GO sheets support continuous water networks essential for high proton conductivity. In ion-exchange membranes, GO contributes both selective ion transport and mechanical reinforcement, as its negatively charged surface can enhance cation selectivity and suppress undesired anion crossover. Furthermore, the adjustable flake size and oxidation degree of GO enable fine control of channel alignment and interlayer spacing, allowing membranes to balance conductivity, selectivity, and durability. As a result, GO-based or GO-modified proton and ion-exchange membranes have shown improved thermal stability, reduced swelling, and superior operational performance compared to traditional polymer systems, positioning GO as a promising material for next-generation energy, electrochemical, and separation technologies. Aixalà-Perelló et al. investigated the scalable fabrication of GO membranes for ion-exchange applications with tunable permselectivity [145]. Their results demonstrated that GO membranes exhibit strong ion selectivity and effective size exclusion toward monovalent cations, achieving permselectivity values of up to 96%. The study also highlighted UV-light irradiation as a promising green reduction strategy, where partial reduction of GO decreased nanochannel dimensions and membrane swelling, thereby further enhancing permselectivity. The incorporation of polymeric binders was explored to improve mechanical robustness; however, while PVP and SPEEK provided performance comparable to self-standing GO, PVA negatively affected transport properties due to pore shrinkage and non-uniform polymer distribution. Impedance spectroscopy measurements revealed an ionic resistance of 4.6 Ω cm² for the GO-20 membrane, with resistance increasing alongside membrane thickness and reduction time. Ionic resistance was shown to be strongly influenced by the electrolyte confined within GO nanochannels and could be further reduced to 3.9 Ω cm² with extended incubation, a value comparable to commercial ion-exchange membranes and nearly four times lower than previously reported GO-based systems. Notably, the membranes achieved an exceptional cation transference number of 0.97 even under a modest fivefold concentration gradient, corresponding to a maximum power density of 0.54 W m⁻². These results collectively underscore the strong potential of GO as a scalable and high-performance material for next-generation ion-exchange membrane technologies [145]. Gahlot et al. developed GO-reinforced nano-composite ion-exchange membranes (IEMs) by incorporating varying GO loadings (0.5–10 wt%) into a sulfonated polyethersulfone (SPES) matrix, producing membranes approximately 180 μm in thickness with significantly enhanced electrochemical performance [146]. The addition of GO effectively modulated the transport properties of SPES. Among all formulations, the composite containing 10 wt% GO exhibited the highest ionic conductivity, along with improved methanol crossover resistance and selectivity. Desalination performance, evaluated through ionic flux, power consumption, and current efficiency, demonstrated that the 10 wt% GO membrane achieved an ionic flux of 3.51 mol m⁻² h⁻¹, a power consumption of 4.3 kWh kg⁻¹, and a current efficiency of 97.4% during salt removal. Compared to pristine SPES, this represents a 19% increase in ionic flux, a 12% increase in current efficiency, and a 20% reduction in power consumption. The strong interfacial interactions facilitated by the GO nanofillers also enhanced the thermal and mechanical stability of the membranes. Overall, the improved performance and robustness of the GO/SPES nanocomposite membranes highlight their strong potential for practical applications such as direct methanol fuel cells (DMFCs) and electrodialysis. Yu et al. reported that competition from interfering cations hinders the practical scale-up of the Donnan dialysis (DD) process for recovering ammonia nitrogen (NH⁴⁺–N) from wastewater [147]. Achieving highly efficient NH⁴⁺ selective permeation through cation exchange membranes (CEMs) requires precise tuning of the membrane’s surface charge and structure. In this study, a novel CEM was developed by forming a GO–PEI cross-linked layer via self-assembly of GO and PEI on the surface of a commercial CEM. This modification strategically adjusts the membrane’s surface charge and structure. The positively charged membrane surface exhibits stronger repulsion toward divalent cations than monovalent ones due to Coulombic interactions, while GO facilitates π–metal cation conjugation (e.g., with Mg²⁺ and Ca²⁺), thereby restricting metal cation transport. In DD experiments, higher NH⁴⁺ concentrations increased the time to reach Donnan equilibrium and enhanced NH⁴⁺ flux, whereas elevated Mg²⁺ concentrations reduced NH⁴⁺ flux (from 0.414 to 0.213 mol·m⁻²·h⁻¹). Leveraging the combined effects of electrostatic repulsion and non-covalent cross-linking, the GO–PEI (20) membrane, prepared by 20 min immersion in the GO–PEI solution, achieved an NH⁴⁺ transport rate of 0.429 mol·m⁻²·h⁻¹, a Mg²⁺ transport rate of 0.003 mol·m⁻²·h⁻¹ in single-salt tests, and an NH⁴⁺/Mg²⁺ selectivity of 15.46, surpassing unmodified and PEI-only membranes (1.30 and 5.74, respectively). In mixed-salt solutions, GO-PEI (20) maintained a high NH⁴⁺/Mg²⁺ selectivity (15.46 vs. ~1.36 for the unmodified membrane) and demonstrated excellent structural stability over 72 h of continuous operation. This facile surface charge modulation strategy thus offers a promising route for enhancing NH⁴⁺-selective separation in CEMs. The proton exchange membrane (PEM) in PEM water electrolyzers is highly susceptible to chemical reactions, mechanical stress, and temperature fluctuations, making it one of the system’s most vulnerable components. Membrane perforation or failure can lead to catastrophic cell damage due to hydrogen and oxygen mixing. Moving beyond conventional casting techniques for reinforced PEMs, Ceballos-Alvarez et al., in their study, explored a novel approach by depositing a composite Nafion^®–GO layer onto an extruded commercial Nafion^® 115 membrane [148]. A GO–Nafion^® ink was formulated, and a 20 μm composite layer was applied via ultrasonic spraying. Various thermal treatments were investigated to enhance bonding between the composite layer and the extruded membrane. This method produced membranes with increased hydrophilicity, greater strain tolerance, and higher tensile load capacity, particularly in samples annealed at elevated temperatures. When applied to the cathode side, the composite layer improved mechanical and thermal stability without compromising electrochemical performance, despite increasing the membrane thickness by over 15%. Deposition on the anode or on both sides, however, reduced electrochemical performance by approximately 9% and 15%, respectively. Overall, all GO–Nafion^® membranes exhibited good long-term stability, highlighting the Nafion^®–GO composite as a promising material for water electrolysis applications. Kumar Varshney et al. reported that GO-reinforced perfluorosulfonic acid (PFSA) proton exchange membranes (PEMs) exhibit enhanced ion diffusion, leading to improved polymer electrolyte fuel cell (PEFC) performance [149]. However, the mechanisms by which GO affects water dynamics and hydronium ion transport remain underexplored. It is also anticipated that the interlayer spacing of multilayer GO plays a key role in facilitating ion mobility. This study investigates water and ion dynamics in GO–PFSA membranes and examines how GO interlayer spacing influences ion diffusion. Molecular dynamics (MD) simulations were employed to analyze the behavior of multilayer GO within the PFSA matrix and to probe interactions between GO surface functional groups (epoxy and hydroxyl) and water molecules and hydronium ions. Water retention near the multilayer GO is shown to be critical in forming transport channels that enhance ion mobility within the membrane. An optimal interlayer spacing of 9.5 Å was identified as the threshold at which ion diffusion reaches its maximum. Compared with pristine Nafion^®, the hydronium ion diffusion coefficient in multilayer GO–PFSA membranes improved by approximately 17% and 30% at 300 K, and 9% and 12% at 350 K, for hydration levels (λ) of 13 and 20, respectively.

6. Critical Assessment

A persistent challenge in the GO research landscape is the significant variability introduced by differences in synthesis routes and post-processing conditions. Although many studies refer broadly to “Hummers” or “Hummers-modified” methods, the practical implementation of these protocols varies widely in terms of oxidizing agents, reaction temperature, reaction time, and purification steps [75,76,150]. These variations influence the oxidation degree, defect density, flake size distribution, and functional group composition of GO, ultimately affecting the reproducibility and comparability of findings across the literature.

GO produced via classical Hummers, modified Hummers, or alternative oxidation routes (e.g., improved Hummers, electrochemical oxidation) exhibits substantial differences in C/O ratios, oxygen functional group distribution, and structural disorder [22,150]. A higher oxidation degree typically leads to increased interlayer spacing and enhanced hydrophilicity, which can improve water permeation but may compromise mechanical stability and chemical robustness. Conversely, partially oxidized or mildly oxidized GO often displays improved mechanical integrity and stability under pressure-driven filtration, but exhibits lower functional-group density and reduced interaction with specific contaminants or ions. These structural differences directly influence transport behavior in GO-based membranes [151]. Highly oxidized GO promotes fast water transport through widened capillaries; however, the associated high density of defects and oxygen functionalities may contribute to swelling, layer delamination, and performance drift over time. Less oxidized GO, on the other hand, forms tighter and more stable nanochannels but may sacrifice permeability for improved selectivity [152,153].

Washing and purification steps—including centrifugation, dialysis, pH adjustment, and filtration—are a major source of batch-to-batch variability. Inadequate removal of residual acids, metal ions (e.g., Mn²⁺, Fe³⁺), or small oxidative fragments can alter the surface charge, ionic conductivity, and chemical stability of GO dispersions. Residual ions may also catalyze undesired reactions, affect membrane crosslinking chemistry, or promote aggregation. Furthermore, differences in final dispersion concentration and ionic strength strongly influence flake stacking behavior, ultimately affecting membrane microstructure and reproducibility [154,155].

Post-processing steps such as thermal annealing, chemical reduction, sonication, and size-fractionation modify the electronic structure and physical morphology of GO. Reduction processes decrease oxygen content and restore sp² conjugation, thereby narrowing interlayer spacing and modifying water transport pathways. However, the extent of reduction varies significantly with reagent type (ascorbic acid, hydrazine, NaBH₄), concentration, pH, treatment duration, and temperature. This introduces substantial variability in membrane permeability, ion selectivity, antifouling properties, and long-term stability. Excessive sonication may produce small, highly defective flakes that form disordered laminar structures with poor mechanical strength, while insufficient exfoliation can yield large, multilayer aggregates that impede uniform membrane formation [156,157,158]. Due to the cumulative influence of the above parameters, GO produced from nominally identical procedures often varies significantly from batch to batch. These differences have notable consequences for membrane fabrication and performance.

Variations in flake size distribution and oxidation degree lead to inconsistent interlayer spacing and nanochannel alignment, affecting permeability, selectivity, and swelling behavior. Differences in defect density and flake integrity influence membrane robustness under pressure. Highly defective or over-oxidized batches may yield membranes prone to cracking, delamination, or compaction. Ionic and molecular transport—regulated by capillary size, surface charge, and hydrophilicity—changes substantially with variations in oxidation level, functional group density, and hydration behavior. Batch-dependent fluctuations in C/O ratio, d–spacing (dry/hydrated), and flake morphology result in inconsistent flux, rejection, and long-term operational stability, limiting direct comparison across studies [153,159,160,161].

Despite their promising performance, GO-based membranes and composites face several critical limitations that constrain their practical deployment. A major concern is swelling instability, particularly in aqueous environments, where hydration-driven expansion of GO interlayer spacing can diminish selectivity and create uncontrolled transport pathways. Similarly, restacking of GO nanosheets, often triggered by van der Waals forces or insufficient dispersion, can collapse nanochannels, reduce permeability, and lead to heterogeneous membrane structures. Long-term operational instability also remains a challenge, as continuous exposure to variable pH, ionic strength, oxidants, or mechanical stress may degrade GO functional groups or weaken polymer–GO interfacial interactions, ultimately compromising membrane durability. In addition, the potential toxicity of GO, including oxidative stress effects on microorganisms and cells, raises concerns for both worker safety and downstream environmental release. These issues highlight the need for improved stabilization strategies, comprehensive lifetime assessments, and environmentally responsible design approaches to fully realize the benefits of GO-based membrane technologies.

Overall, the significant variability arising from synthesis, purification, and post-processing steps underscores the need for standardized reporting practices and minimum characterization criteria. Without such standardization, the reproducibility and comparability of GO-based membrane performance remain limited, hindering progress toward scalable and industrially relevant applications.

Finally, GO-based membranes and GO–polymer composites generally outperform pristine polymer membranes due to the unique structural and interfacial characteristics of GO. The presence of 2D GO sheets introduces well-defined, hydrophilic nanochannels that enhance water permeability while maintaining or improving solute selectivity through controlled interlayer spacing and steric hindrance. In addition, the abundant oxygenated functional groups on GO participate in strong hydrogen bonding, π–π interactions, and electrostatic forces with polymer chains, creating a reinforced interphase that restricts chain mobility and thereby improves mechanical strength and thermal stability. These interactions also increase the tortuosity of diffusion pathways, contributing to reduced fouling and improved barrier properties. Collectively, these synergistic effects explain why GO-containing membranes exhibit superior transport performance, stability, and durability compared to their pure polymer counterparts.

7. Challenges and Limitations

Despite the rapid growth of GO membrane research and the impressive performance demonstrated in laboratory-scale studies, several key limitations continue to constrain their transition from experimental prototypes to industrially deployable technologies. These limitations span structural durability, fabrication scalability, long-term operational stability, environmental impacts, and economic feasibility. To better align with applied research needs, this section provides an expanded discussion incorporating quantitative benchmarking from the literature, techno-economic perspectives, and recent advances directed at improving scalability and sustainability.

• Structural Stability, Mechanical Integrity, and Operational Lifespan—GO membranes remain susceptible to structural degradation under realistic operating conditions. Prolonged exposure to hydraulic pressure (>3–5 bar), transmembrane shear, or chemically aggressive feed streams can induce layer slippage, interlayer spacing fluctuations, or partial delamination due to the relatively weak van der Waals interactions holding GO lamellae together. Reported operational lifetimes for pristine GO laminates range from several hours to a few weeks, depending on feed chemistry and mechanical support. Stabilization techniques—such as ionic crosslinking, covalent bridging, or polymer intercalation—extend the lifespan to several months, but these strategies often increase fabrication complexity and reduce permeability. A systematic evaluation of long-term performance remains scarce, and only a limited number of studies report durability tests beyond 100–500 h, underscoring the need for standardized lifetime benchmarking [110].
• Transport Performance and the Permeability–Selectivity Trade-off—Achieving simultaneous high permeability and tight molecular/ionic selectivity remains a significant challenge. Minor variations in interlayer spacing (0.1–0.2 nm), defect density, GO oxidation state, or sheet size distribution can shift rejection rates by 10–30%, complicating reproducibility across batches and laboratories. While Section 3.1 addressed the theoretical basis of GO transport mechanisms, Section 4 illustrated these limitations through literature case studies; the combined evidence highlights that transport is strongly modulated by synthesis-induced structural heterogeneities. Emerging strategies—such as 2D heterostructure, controlled nanochannel alignment, and chemical crosslinking—offer promising routes to narrow variability, but require more rigorous quantitative validation and standardized reporting [162,163,164].
• Chemical and thermal stability—GO’s oxygen-containing functional groups, while beneficial for hydrophilicity and tunability, are chemically unstable in extreme pH or high-temperature environments. Reduction or oxidation reactions can modify the membrane chemistry, leading to performance degradation over time. Ensuring long-term chemical resistance without sacrificing tunability remains an active area of research [165].
• Scalability, Production Throughput, and Manufacturing Constraints—Most current GO membrane preparation techniques, including vacuum filtration, drop-casting, spin-coating, and layer-by-layer assembly, remain limited to laboratory-scale throughputs (<0.1–0.5 m²·day⁻¹). Recent advances such as slot-die coating, spray-lamination, roll-to-roll deposition, and continuous shear-alignment methods have demonstrated production rates on the order of tens of m²·day⁻¹, approaching pilot-scale requirements. However, maintaining uniform thickness (<100 nm variation), precise stacking orientation, and defect-free morphology at high throughput remains challenging. Batch-to-batch variations in GO precursor quality—especially sheet size distribution, oxidative degree, and residual metal content—further complicate scale-up. Techno-economic analyses (TEA) are still sparse, but preliminary models estimate that high-quality GO production contributes 40–70% of total membrane cost, with oxidant consumption, washing demands, and waste treatment representing major cost drivers. Current estimates for pilot-scale GO membrane fabrication range from 50 to 200 USD·m⁻², substantially higher than commercial polymeric membranes (5–30 USD·m⁻²), though costs are projected to decline with process intensification and greener synthesis routes [145,166,167].
• Fouling Resistance, Long-Term Operation, and Maintenance—Although the hydrophilic and negatively charged GO surface inherently reduces organic fouling, long-term studies reveal gradual performance deterioration during filtration of complex natural waters, brines, or industrial effluents. Flux decline of 20–60% over 24–72 h has been reported for pristine GO laminates. Biofouling remains particularly problematic, as microorganisms can colonize oxygenated functional groups. Antifouling strategies such as surface grafting, zwitterionic modification, and silver or photocatalytic nanoparticle incorporation show promise but introduce concerns regarding toxicity, leaching, and long-term stability. Few studies quantify cleaning cycles, recovery ratios, or maintenance intervals, critical metrics for evaluating operational costs and membrane lifetime [167,168,169].
• Environmental and Toxicological Considerations—Conventional GO synthesis relies on strong oxidizing agents (KMnO₄, concentrated H₂SO₄, NaNO₃), producing metal-containing acidic waste that requires intensive neutralization and treatment. Estimates suggest that 20–40 L of acidic wastewater can be generated per gram of purified GO using unoptimized Hummers-type methods. Life-cycle assessments (LCAs) indicate that energy consumption during oxidation, exfoliation, and multi-stage washing dominates the environmental footprint, while concerns regarding the ecotoxicity of released GO nanosheets—particularly their interactions with microorganisms and aquatic species—remain unresolved. Recent developments, including electrochemical exfoliation, green oxidants (e.g., persulfates, H₂O₂-assisted oxidation), solvent-free or low-acid methods, and improved waste recovery, offer compelling routes toward more sustainable GO production. However, scalability and consistency remain technological hurdles [170,171].

Overall, while GO membranes hold considerable promise, their translation into commercial separation technologies is constrained by interconnected limitations in mechanical durability, large-scale manufacturing, environmental sustainability, and economic competitiveness. Addressing these gaps requires coordinated efforts toward standardized testing protocols, comparative benchmarking against commercial membranes, large-scale durability studies, and the development of greener, economically viable GO production pathways. Comparative analyses provided below (

Table 2. GO membranes vs. commercial polymeric membranes: quantitative benchmarking.

Parameter	GO Membranes (Typical Lab-Scale)	Commercial Polyamide RO/NF	Ceramic/Alumina Membranes
Water permeability	10–60 L·m−2·h−1·bar−1	1–5 L·m−2·h−1·bar−1	1–5 L·m−2·h−1·bar−1
NaCl rejection	20–85%	>95%, 40–90%	>95%, 40–90%
Operational pressure	1–5 bar	10–70 bar	10–70 bar
Typical lifespan	Weeks–months	3–7 years	3–7 years
Fouling resistance	Moderate; improves with modifications	Moderate	Moderate
Typical cost (USD·m−2)	50–200	5–30	5–30

) outline the current state of GO membrane performance relative to established technologies and highlight areas requiring further research.

Additional studies are required to further refine the membrane’s performance (Figure 13).

Addressing these challenges requires interdisciplinary efforts combining materials engineering, surface chemistry, and process optimization to develop robust, cost-effective, and environmentally sustainable GO membrane systems.