Melamine-Functionalized Graphene Oxide as a Multifunctional Modifier for High-Performance Epoxy Nanocomposites with Enhanced Mechanical Properties and Thermal Stability

TL;DR

This paper introduces a method to enhance epoxy composites by functionalizing graphene oxide with melamine, resulting in improved mechanical strength and thermal stability.

Contribution

The novel approach of covalently functionalizing graphene oxide with melamine to create high-performance epoxy nanocomposites is presented.

Findings

Melamine-functionalized graphene oxide reduces gelation time and increases self-heating temperature during curing.

The nanocomposites show significant improvements in tensile, flexural, and impact strengths, as well as hardness.

Thermal stability metrics like T5%, T50%, and THRI show substantial increases, confirming the formation of a thermally stabilizing barrier.

Abstract

Developing polymer composites with improved mechanical and thermal properties requires strategies to overcome the problem of agglomeration and weak interfacial interactions of carbon nanofillers. This paper presents an effective strategy for the covalent functionalization of graphene oxide (GO) with melamine to create high-performance epoxy nanocomposites. The functionalization results in the formation of nitrogen-containing heterocyclic structures on the GO surface, as confirmed by FTIR and Raman spectroscopy. The addition of the obtained modified filler (mel-GO) into the epoxy matrix provides a synergistic effect: the melamine amino groups catalytically accelerate curing, reducing the gelation time from 146 to 48 min and increasing the maximum self-heating temperature from 94 to 122 °C, thus indicating enhanced interfacial interaction. This interaction results in a remarkable overall improvement in mechanical properties: tensile and flexural strengths increase by more than 20%, and elastic moduli by 31% and 58%, respectively, compared to the composite containing unmodified GO. At the same time, impact strength (from 14 to 23 kJ/m2) and hardness (up to 87 Shore D) increase. A key achievement is a dramatic increase in thermal and thermal-oxidative stability: the onset temperature of decomposition (T5%) increases by 27 °C, the half-decomposition temperature (T50%) by 45 °C, and the thermal stability index (THRI) increases from 119.3 to 128.9 °C. A more than twofold increase in coke residue yield (to 9.29%) and an increase in the Vicat softening point to 175 °C confirm the formation of an effective thermally stabilizing barrier layer due to the combined action of nitrogen-containing groups and dispersed graphene flakes. The proposed approach to functionalizing graphene oxide with melamine opens the way for the creation of next-generation epoxy composites with a record-breaking combination of strength, impact toughness, and thermal stability for applications in aerospace, electronics, and composite structures operating under extreme conditions.

Figures

Full text

1. Introduction

The ambitious goals facing modern science, caused by global challenges and needs of the society, require the development of new materials with a specific set of unique properties. Such materials traditionally include composites, nanomaterials, smart materials, metamaterials, biomaterials, and others. Polymer composite materials, which have proven themselves in various industries, occupy a special place among them. Their key advantage is a wide range of polymers, which, combined with the ability to be specifically modified, allows us to select the structure of the material for specific operating conditions [1,2]. Epoxy resins have found the most widespread application among thermosetting plastics. However, currently produced high-volume grades exhibit inherent limitations in strength, rigidity, thermal stability, and flame retardancy, necessitating effective modification strategies to impart new functional properties. As documented in recent comprehensive reviews, the optimization of mechanical performance [1], incorporation of graphene-based multifunctional modifiers [2], advancements in thermal management applications [3], understanding of nanofiller effects on thermal properties [4], and development of flame-retardant epoxy nanocomposites [5] represent key research directions addressing these challenges. These considerations underscore the need for continued development of novel approaches to overcome the inherent shortcomings of epoxy materials.

Current approaches to producing modern epoxy composites involve not only the use of dispersed and fibrous fillers but also the application of various types of nano-materials [6,7], functionalizing agents, external energy inputs, thermoplastic modifiers, and the creation of vitrimers [8] and hybrid [9] systems with complex architectures [10,11]. Functionalized nanoparticles are of particular interest because they combine the advantages of nanofilling and chemical modification of the matrix: grafting functional groups onto the highly developed surface of nanoscale particles enhances adhesion at the interface, reduces the tendency of particles to agglomerate, and, in some cases, allows us to impart additional functional properties to the composite. The most widely used are carbon nanomaterials: single- and multi-walled carbon nanotubes (CNTs, MWCNTs) [12,13], graphene oxide (GO) [14,15], reduced GO [16,17], nanodiamonds [18,19], etc. Numerous studies show high potential of using GO as a modifying agent for an epoxy matrix, since it is characterized by good dispersibility and comparative ease of functionalization by various chemical approaches, which opens up the possibility of constructing composites with a given set of properties. The studies presented in [14] on the modification of fiber-reinforced epoxy composites with nanoparticles of multi-walled carbon nanotubes (MWCNTs), graphene nanoplatelets (GNPs), and GO indicate that the best combination of properties is achieved with a GO content of 0.4 wt.%. At this concentration, the tensile strength and toughness (area under the stress–strain curve) of the composite increase by 49 and 160%, respectively, compared to the pristine unmodified material. Dynamic mechanical analysis (DMA) showed an increase in the storage modulus by 31% and an increase in the glass transition temperature from 49 to 54 °C, which the authors attribute to the chemical interaction at the GO–epoxy matrix interface. The study [15] showed that with an increase in the GO content to 0.2 wt.%, the tensile strength and elastic modulus increased by 16 and 17%, respectively. The best value of Poisson’s ratio (ν = 0.38) was recorded at the GO content of 0.1 wt.%, while the maximum storage modulus was achieved with the addition of 0.2 wt.% GO. A tendency towards a decrease in the glass transition temperature from 78.7 to 64.2 °C with an increase in the filler content was noted. Moreover, only a slight increase in the heat-resistance index from 157.8 to 159.4 °C was observed, indicating a weak increase in the thermal stability of the material. Wei et al. [20] found that the degree of GO oxidation is a critical parameter determining the balance between good dispersion of the nanofiller and effective interfacial interaction. The GO-4 sample with an intermediate oxidation state increased the crack resistance of the epoxy composite by 128% compared to pure epoxy, which is likely due to a smaller number of agglomerates and the chemical interaction of the –OH and –COOH functional groups with the matrix. A review of the presented studies demonstrates that the effect of modifying epoxy systems with GO is complex, parameter-sensitive, and is determined not only by the concentration of the nanofiller but also by the nature of its surface. Therefore, further development in this field requires a systematic study of approaches to GO functionalization and the selection of surface groups that provide an optimal balance between dispersibility and interfacial interactions. Study [21] showed that the functionalization of GO with bis-furan di-epoxide (bGO) helps reduce the amount of agglomerates due to the improved interfacial interaction with the epoxy matrix. With the addition of 0.5 wt.% bGO, the tensile strength increased by 80%, while the flexural strength and impact toughness increased by 49 and 130%, respectively. A 97% increase in crack growth energy was noted, and the storage modulus and glass transition temperature increased with increasing bGO content (from 90.3 to 116.8 °C). The authors attribute the observed strengthening of the matrix to the uniform distribution of bGO and strong interfacial adhesion, which ensures efficient load transfer and deflection of developing cracks, leading to a simultaneous increase in the modulus, strength, and crack resistance of the material. Rhili et al. [22] showed that functionalization of GO with hexachlorocyclotriphosphazene and p-phenylenediamine resulted in an increase in coke yield and a decrease in the amount of gaseous pyrolysis products. According to the authors, this change in the decomposition ways promotes the formation of a more stable protective layer, allowing the modified polymer to self-extinguish after the flame source is removed. It has been shown [23] that functionalization of GO with a tertiary amine synthesized from diethanolamine and cardanol epoxy resin (GOF) leads to the increased thermal stability of the composite, as well as the reduction in the coefficient of friction from 0.567 to 0.408.

An analysis of the literature data clearly demonstrates that targeted modification of the chemical structure of the graphene oxide (GO) surface is a key tool for controlling the functional and performance properties of epoxy composites. In this context, melamine is of particular interest as a functionalizing agent. Its molecular structure, containing reactive amine and triazine groups, provides a dual advantage. First, these groups are capable of specific interactions with the carbon surface of GO, facilitating its stable and uniform dispersion in the polymer matrix and reducing its tendency to aggregation [10,24,25]. Secondly, these same functional groups can act as active centers, chemically interacting with the components of the epoxy system, directly enhancing interfacial adhesion and load transfer efficiency. This combination of effects creates the fundamental prerequisites for fully realizing the reinforcing and barrier potential of the two-dimensional nanofiller. Furthermore, melamine itself is known as a functional nitrogen-containing additive capable of modifying thermal decomposition and increasing the flame retardancy of polymers, creating additional opportunities for imparting desired properties to epoxy composites and further justifying the relevance of melamine-mediated modification strategies [26,27].

2. Materials and Methods

2.1. The Materials and Reagents

ED-20 epoxy resin (epoxy group content 20–22%, per supplier datasheet) and PEPA hardener (primary amine content 28–32%, per supplier datasheet) were purchased from CHIMEX Ltd. (St. Petersburg, Russia) were used without further purification. Tris(1-chloro-2-propyl) phosphate (TCPP) from Xuancheng City Trooyawn Refined Chemical Industry Co., Ltd. (Xuancheng, China) was used as a plasticizer and a fire retardant. Melamine (2,4,6-triamino-1,3,5-triazine; CAS 108-78-1; Sigma-Aldrich (Shanghai, China), product no. 52549) was used as received. GO was synthesized by galvanostatic anodic oxidation of dispersed natural graphite in 83 wt. % H_2_SO_4_ (200 mA·g^−1^; specific charge 700 mA·h·g^−1^), followed by hydrolysis for 15 min at 15–18 °C and drying at 90–100 °C, as described previously. The yield of GO after synthesis was 160%. This apparent yield exceeding 100% is typical for GO synthesis and results from the incorporation of oxygen-containing functional groups (epoxy, hydroxyl, carboxyl) and intercalated water molecules into the graphite structure during oxidation and subsequent hydrolysis [28,29]. It should be noted that quantitative determination of oxygen-containing functional groups (particularly carboxyl groups) by classical titration methods is challenging for electrochemically synthesized GO. This is due to the inaccessibility of active sites within aggregated structures, the instability of GO in alkaline media, and surface charge alterations during titration [28]. A comprehensive analysis would require complementary techniques such as XPS, which is not currently available in our laboratories. Therefore, the present study relies on spectroscopic (FTIR, Raman) and diffraction (XRD) methods to confirm successful functionalization, while the quantitative determination of specific functional groups remains a direction for future investigation.

2.2. Obtaining Functionalized GO

GO dispersion (0.05 g) was prepared in 150 mL of deionized water using an Stegler 3DT ultrasonic bath (Taipei, Taiwan (China)) operating at a frequency of 40 kHz for 1 h to ensure complete exfoliation and uniform dispersion of graphene oxide sheets. Melamine (0.5 g) was then added and stirred on a magnetic stirrer at 25 °C for 24 h. The resulting suspension was centrifuged twice at 8000 rpm for 20 min per cycle. After the first centrifugation, the supernatant was discarded, and the sediment was re-dispersed in fresh deionized water for washing to remove unreacted melamine and soluble by-products, followed by a second centrifugation under the same conditions. The purified product was then dried at 90 °C in air to constant weight. The dried powder was placed in a corundum crucible and annealed at 600 °C (heating rate 10 °C/min) for 1 h, Figure 1.

2.3. Characterization of the GO

The structural and chemical modifications of GO were characterized using complementary techniques to establish structure-property relationships. FTIR and Raman spectroscopy were employed to identify covalent functionalization and carbon lattice modifications, as these methods are highly sensitive to chemical bonding and structural defects in carbon nanomaterials [30,31]. XRD analysis was performed to assess interlayer spacing changes and crystalline structure evolution upon functionalization [32,33]. SEM provided direct visualization of morphology and elemental composition, essential for understanding filler distribution and surface chemistry [34,35].

X-ray diffraction (XRD) patterns for GO and mel-GO were recorded on an ARL X’TRA diffractometer (Thermo Scientific, Ecublens, Switzerland) using CuKα radiation (λ = 0.15412 nm). The images were collected in the range of 2θ = 5–50° at a scan rate of 2°/min.

The surface morphology and elemental composition of GO and mel-GO were examined using scanning electron microscopy with energy-dispersive analysis (SEM/EDS) on an ASPEX EXplorer™ instrument (Pittsburgh, PA, USA).

Fourier transform infrared (FTIR) spectra were recorded using a Simex spectrometer (Novosibirsk, Russia) in the 4000–500 cm^−1^ range at room temperature in ATR mode. Spectra were recorded at a resolution of 4 cm^−1^, accumulating 45 scans per spectrum.

Raman spectra were recorded using a Thermo Fisher Scientific spectrometer (Waltham, MA, USA) equipped with a confocal microscope using a 532 nm laser.

2.4. Preparation of Epoxy Composites

TCPP (40 parts by weight) and a specified amount (0.1 parts by weight) of GO/mel-GO were added into the pristine resin. The component ratios (epoxy:TCPP:PEPA = 100:40:15 by weight) and the nanofiller content (0.1 parts by weight) were selected based on our previous optimization studies [36,37], which established the formulation providing optimal processability and performance. To ensure direct comparability between the unmodified GO and mel-GO systems, identical compositions were used for both nanofillers. Dispersion was performed using a probe ultrasonic disperser (nominal operating frequency of 22 kHz, nominal generator output power of 400 W). Ultrasonic treatment was carried out by cooling with a water bath with running water (15–20 °C) for 60 min; the composition temperature during treatment did not exceed 50 °C. Next, PEPA hardener (15 parts by weight) was added and mixed at 100 rpm for 3 min. The resulting composition was poured into silicone molds and kept at room temperature for 72 h, after which post-curing was carried out in two stages: for 8 h at 90 °C (during the first 24 h) and for 8 h at 120 °C.

2.5. Composite Testing

Mechanical testing methods were selected based on their relevance to structural applications and established international standards. Tensile and flexural tests provide fundamental parameters—strength, modulus, and deformation behavior—that are essential for engineering design and material comparison. Impact testing evaluates fracture toughness under dynamic loading, a critical property for composites subjected to sudden stress. Shore D hardness offers a rapid, non-destructive measure of surface resistance to indentation, complementing the bulk mechanical properties [2,38].

Tensile tests were performed on dog bone specimens (4 mm thick, 10 mm wide, 50 mm of working part length), and bending tests were performed on block specimens (4 × 10 × 80 mm). Impact toughness was determined on block specimens (15 × 10 × 120 mm). Tensile and bending tests were performed using a WDW 5E universal testing machine (Time Group Inc., Beijing, China) at a crosshead speed of 5 mm/min and 50 mm/min, respectively, in accordance with the standards [39,40]. Impact tests were performed using a LCT-50D device (Beijing United Test Co., Ltd., Beijing, China) in accordance with the standard [41]. The hardness of the specimens was determined according to Shore D using a PCE-DD-D Shore D hardness tester (PCE Instruments UK Ltd., Manchester, UK) in accordance with the standard [42]. All mechanical tests were conducted under controlled environmental conditions: temperature of 23 ± 2 °C and relative humidity of 50 ± 10%. For each composition, 5 specimens (n = 5) were tested. Before testing, the specimens were conditioned for 72 h at 23 ± 2 °C and subjected to heat treatment according to the post-curing regime described in Section 2.4.

The Vicat softening point was determined using the B50 method (50 N load, 50 °C/h heating rate). The polymerization process of the composites was studied by recording the self-heating temperature over time after the addition of the hardener. For this purpose, the composite was placed in an insulated cell equipped with a thermocouple installed in the center of the sample. Temperature changes were recorded in real time immediately after the addition of the hardener and the start of mixing. Tests were conducted at a controlled ambient temperature of 25 °C provided by a thermostatted chamber, which eliminated external thermal effects. The resulting temperature–time dependence was used to analyze heat generation, evaluate the intensity of the curing reaction and compare the polymerization kinetics of various composites.

Thermal analysis methods were selected to investigate both processing behavior and service performance. DSC was employed to study curing kinetics and reaction enthalpy, providing insight into how nanofiller functionalization affects network formation [43]. TGA was used to evaluate thermal and thermo-oxidative stability, with characteristic decomposition temperatures and char yield serving as key indicators of material performance at elevated temperatures [44,45]. Both methods are widely used in polymer nanocomposite research and enable direct comparison with literature data. Curing kinetics was studied using the DSC method on a DTAS-1300 instrument (Samara, Russia) on 20 mg samples heated to 250 °C at a rate of 8 °C/min in air. The measurements were conducted in static air atmosphere using standard aluminum crucibles (20–40 μL capacity) with flat bottoms to ensure optimal thermal contact with the DSC sensor. The crucibles were used without lids to maintain consistent atmospheric conditions with the DSC experiments. TGA was performed using a Q-1500D derivatograph (MOM, Budapest, Hungary) on 100 mg samples by heating to 900 °C at a rate of 10 °C/min in air. The experiments were conducted in static air atmosphere without forced gas flow, using ceramic alumina (Al_2_O_3_) crucibles. The crucibles had a capacity of approximately 100 mg and were used without lids to allow direct contact with the oxidizing atmosphere.

3. Results and Discussion

3.1. Mechanical Properties of Epoxy Composite

Figure 2 shows test results of tensile and flexural strength for epoxy composites containing GO and mel-GO. The addition of GO increased tensile and flexural strength by 52.9% and 17.7%, while the corresponding elastic moduli increased by 14.9% and 31.1%. Melamine functionalization of GO resulted in an increase in the stiffness and strength of the composite. Tensile and flexural strength increased by more than 20%, and elastic moduli by 31% and 58%, respectively, compared to the GO-modified composite.

The addition of GO (Figure 3) leads to an increase in the impact strength of the composite from 9 to 14 kJ/m^2^, while grafting melamine onto the GO surface further increases this value to 23 kJ/m^2^. A similar trend is observed for hardness: its value increases from 76 to 87 units.

An overall increase in strength, elastic modulus, and hardness with the addition of GO is consistent with the effective reinforcement of the epoxy matrix and more efficient load transfer to the nanofiller. For melamine-modified GO (Figure 2 and Figure 3), an additional increase in the entire range of properties is observed, which is likely due to the combination of factors. Specifically, an increase in strength may be due to a reduction in the defectiveness of the composite structure, including a decrease in the number or size of filler agglomerates, which reduces local stress concentration. The simultaneous increase in elastic modulus and impact strength indicates the formation of a more rigid structure without loss of energy dissipation capacity under impact loading, which is indirectly consistent with a more uniform particle distribution within the matrix and/or enhanced interfacial interactions [46,47,48].

The observed simultaneous enhancement of stiffness (elastic modulus) and toughness (impact strength) upon mel-GO incorporation is a critical finding that merits mechanistic explanation. In classical polymer mechanics, increasing stiffness often leads to embrittlement due to restricted chain mobility and reduced plastic deformation. However, in carbon-based polymer nanocomposites, a strong and well-engineered interface can decouple these properties [46,47,48]. The covalently bonded melamine functionalities on the GO surface (confirmed by FTIR and Raman in Section 3.5) play a dual role. Firstly, they establish robust interfacial adhesion, enabling efficient stress transfer from the matrix to the high-modulus GO filler, which increases the overall composite modulus. Secondly, this strong interface facilitates energy dissipation through mechanisms other than bulk matrix deformation. As cracks initiate and propagate, the chemically bonded mel-GO sheets promote crack deflection and crack bridging, forcing the crack to travel a longer, more tortuous path, which absorbs more energy [49,50,51,52,53]. Furthermore, the nitrogeneous groups on mel-GO may create an interphase region with a gradient of properties, allowing for local yielding without macroscopic failure. Similar synergistic effects, where reinforcing elements improve both strength and ductility through effective crack control and interfacial engineering, have been observed in hybrid composite systems [54,55]. Thus, the stiffness–toughness balance achieved here is not a contradiction but a direct consequence of the superior interfacial design enabled by melamine functionalization.

3.2. Fractography Analysis of the Obtained Epoxy Composites

Figure 4 presents the fractography data for the pristine epoxy composite (Figure 4a), modified with GO (Figure 4b) and modified with mel-GO (Figure 4c,d). As can be seen in Figure 4a, the pristine epoxy composite is characterized by a smooth cleavage surface, indicating its low ability to resist crack formation and propagation under loading, which is consistent with a predominantly brittle fracture mechanism. The addition of GO into the epoxy matrix, Figure 4b, led to the formation of a pronounced stepped-layered relief during fracture, indicating a more tortuous crack trajectory and an increase in the fracture surface area [49,50,51]. Melamine treatment of GO (Figure 4c,d) is accompanied by the increased development of the fracture surface relief, as well as the appearance of localized cavities and areas of detachment at the matrix–filler interface. The observed changes in the fracture structure may be related to a more uniform particle distribution within the matrix, which increases the number of obstacles to crack propagation, as well as to increased interfacial interactions: in this case, fracture requires additional energy expended on interfacial detachment of mel-GO particles and localized deformation of the surrounding matrix [52,53].

A gradual increase in the number of fracture surface inhomogeneities in the composite chip structure as a result of the addition of GO and mel-GO explains improved mechanical properties due to a change in the nature of the fracture from a predominantly brittle fracture mechanism to a ductile-brittle one, accompanied by crack deflection and branching, the formation of a stepped developed relief and, probably, local interphase separation at the matrix–filler boundary, which increases the energy required for crack propagation.

3.3. Analysis of the Curing Process of Epoxy Compositions

The observed differences in mechanical properties and fracture surface morphology may be related not only to the distribution of the nanofiller but also to the peculiarities of the network structure formation during curing. Therefore, the curing process of epoxy composites was studied using self-heating temperature monitoring and differential scanning calorimetry to assess the effect of GO and mel-GO on the structure formation of the epoxy matrix.

As can be seen in Figure 5, the addition of GO leads to a slowdown in the curing reaction, which is shown in an increase in the gelation time from 104 to 146 min and an increase in the total curing time from 146 to 195 min, Table 1. At the same time, the maximum self-heating temperature increases from 88 to 94 °C. The DSC data confirm the tendency for the onset of curing to shift to higher temperatures, as the onset temperature of the curing reaction increases from 59 to 71 °C, and the specific heat of the reaction decreases from 536 to 446 J/g, Table 2.

In contrast to the composition with GO, the composition with mel-GO demonstrates a significantly earlier and more intense heat release during self-heating: the gelation time is reduced to 48 min, the curing time to 74 min, and the maximum curing temperature increases to 122 °C, Table 1, Figure 5. According to the DSC data for the system with mel-GO, the onset of curing also shifts to the region of lower temperatures, so the onset of the reaction corresponds to 52 °C relative to 71 °C for unmodified GO, which is also accompanied by an increase in the enthalpy of the reaction from 446 to 507 J/g, Table 2, Figure 6.

The observed slowdown in polymerization processes in the presence of GO may be due to an increase in the viscosity of the composition and diffusion limitations, as well as the interaction of the polar surface of GO with the curing agent components, which reduces the effective mobility of the reactants in the early stages [56,57]. Conversely, a sharp acceleration of the process and an increase in the maximum self-heating temperature for the epoxy composition modified with mel-GO may indicate a change in the surface chemistry of the filler and the appearance of additional active interaction sites that facilitate the epoxy groups opening. Furthermore, if the functionalization leads to a more uniform distribution of particles in the matrix, the available interfacial area and the number of potentially active sites increase, which may also explain the observed effect [43,58].

3.4. Analysis of TGA Data of Epoxy Composites

Analysis of the thermogravimetric data and the calculation of the thermal stability index (THRI) [15,44] made it possible to quantitatively compare the influence of GO and mel-GO on the thermal stability of epoxy composites and the nature of pyrolytic decomposition, Figure 7. The observed changes in the temperature characteristics of decomposition and the yield of coke residue can be associated with the barrier effect of the nanofiller and interphase interactions, which, in turn, correlate with the morphology and features of the formation of the composite structure, discussed earlier.

For the pristine epoxy composite, the onset of noticeable mass loss T_5%_ is observed at 190 °C, which corresponds to the initial stage of thermal degradation with the release of volatile products. The temperature corresponding to 50% mass loss is 385 °C, which characterizes the main region of decomposition of the three-dimensional cross-linked structure of the epoxy composite. At 900 °C, the residue yield is 4.45%, indicating the formation of a small amount of carbon residue (Table 3, Figure 7). The THRI for the initial matrix, calculated from TGA data, is 119.3 °C and is used below as a baseline for comparison.

The addition of GO has a minimal effect on thermal degradation processes: a slight positive shift in T_5%_ to 195 °C and T_50%_ to 392 °C is observed, with a slight increase in the coke residue to 4.85% and THRI to 120.8 °C. The limited effect indicates a weak implementation of the potential barrier mechanism of the lamellar filler, which requires the formation of a sufficiently uniform and extended barrier architecture within the matrix [45,59,60]. This is indirectly consistent with the curing data, as the composition with GO shows a shift in the reaction onset to higher temperatures according to DSC and a noticeable increase in the characteristic times of the self-heating process, indicating a change in the early stages of the structure formation, which may result in an inability to form an effective barrier structure.

For the composite containing mel-GO, a consistent shift in the TGA curve toward higher temperatures is observed (Figure 7), with T_5%_ increasing to 217 °C, T_50%_ to 430 °C, and T_60%_ to 498 °C, respectively, relative to the initial matrix (Table 3). The THRI also increases significantly: for the composite containing mel-GO it reaches 128.9 °C, corresponding to an increase of 9.6 °C relative to the initial matrix. Since THRI is calculated from the T_5%_ and T_30%_ temperatures, i.e., characterizes stability in the early stages of thermal degradation, its significant increase confirms an increase in the “threshold” of thermal-oxidative instability and an improvement in resistance to the initial stages of decomposition, which is consistent with a shift in T_5%_ and the entire thermogram as a whole.

Of particular interest is a more than twofold increase in the residue yield at 900 °C—to 9.29% compared to 4.45% for the initial system. This significant increase in carbonized residue reflects a change in the nature of thermal-oxidative degradation and increased stability of the forming coke residue under oxidizing conditions, rather than the addition of a thermally stable carbon component.

The observed effect of mel-GO may be due to a combination of factors. First, the reduction in the number of agglomerates, as well as the improved distribution of the lamellar filler in the matrix is likely to increase the effectiveness of the barrier mechanism, hindering oxygen diffusion and the release of volatile products, which slows down thermal-oxidative degradation and increases the characteristic decomposition temperatures. Second, the surface modification with mel-GO can enhance interfacial interactions and influence the formation of a protective layer upon heating, which is consistent with an increase in residue at 900 °C. These assumptions are indirectly supported by the fact that for the system with mel-OG, more intense heat release during curing and an earlier reaction onset according to DSC, as well as a more developed fracture surface morphology according to SEM data, were previously recorded, which indicates a change in the structure formation and interphase organization of the composite compared to the composition containing unmodified OG, Figure 4 and Figure 5, Table 1 and Table 2.

Moreover, taking into account the nature of the modification, one cannot exclude the contribution of nitrogen-containing surface fragments of mel-GO to an increase in the yield and stability of the residue during thermal-oxidative decomposition, for example, through a change in the ways of the decomposition and stabilization of the forming protective layer.

The thermomechanical thermal stability of the composites, assessed using the Vicat softening point (T_v_), shows the same trend as the thermal-oxidative stability values measured by TGA. For the pristine matrix, T_v_ is 114 °C; the addition of GO increases it to 142 °C, while the use of mel-GO leads to a further increase to 175 °C (Table 3). Thus, modifying GO with melamine improves not only the resistance to thermal-oxidative degradation but also the resistance to thermo-mechanical softening, which is consistent with the increase in elastic modulus and THRI.

3.5. Structure and Chemistry of GO and Mel-GO

Figure 8 shows the X-ray diffraction patterns of GO and mel-GO. GO exhibits a characteristic low-angle reflection in the region of ~10–12° (2θ), corresponding to the increased interlayer distance of oxidized graphene. After the modification, the low-angle signal is significantly weakened, while the X-ray diffraction pattern of mel-GO exhibits a maximum at 2θ = 27.04° and a reflection at 2θ = 43.51°, characteristic of a more graphite-like order. These changes indicate a decrease in the interlayer distance and an increase in the proportion of ordered sp^2^- domains (partial restoration/ordering of the carbon structure) [32,33]. The observed rearrangement of the filler structure may contribute to a more efficient implementation of the barrier effect and an increase in the thermooxidative stability of the composite, which is consistent with the TGA results, as well as with an increase in rigidity and strength for the system with mel-GO.

The FTIR spectrum of melamine-modified graphene oxide (mel-GO) (Figure 9) shows additional bands in the 2200–2260 cm^−1^ region compared to the pristine GO. These bands can be attributed to C≡N vibrations and indicate the formation of N-containing conjugated fragments. Furthermore, the spectrum of mel-GO contains bands corresponding to C–N vibrations (≈1550, 1435, and 1378 cm^−1^), as well as deformation vibrations characteristic of the triazine ring (~810 cm^−1^). At the same time, bands typical of GO (–OH, C=O/C–O–C) are retained, indicating a partially functionalized, but not fully reduced structure [61,62,63].

Based on Raman spectroscopy data (Figure 10), a comparative analysis of the structural changes in GO after its functionalization with melamine (mel-GO) was performed. The spectra of both samples exhibit characteristic D (~1346 cm^−1^) and G (~1588 cm^−1^) bands, which are responsible for the vibrations of sp^2^-hybridized carbon atoms in the hexagonal lattice and the presence of structural defects (grain boundaries, vacancies, heteroatoms), respectively [30,31]. For the melamine-modified sample, the calculated intensity ratio Iᴅ/Iɢ is 1.02, which is higher than that for the pristine GO (0.91). An increase in this ratio indicates increased imperfections in the carbon framework, which may be due to the covalent attachment of melamine molecules to the surface of the graphene flakes, disrupting the periodicity of the sp^2^ network and creating new defects of the sp^3^ carbon type. This increase in the ID/IG ratio can be explained by the following mechanism: during functionalization, the amino groups of melamine undergo nucleophilic attack on epoxide and carboxyl groups present on the GO surface, forming covalent C–N–C and amide linkages [61,62]. This covalent attachment disrupts the local π-conjugation of the graphene network, converting carbon atoms at the bonding sites from planar sp^2^ hybridization to tetrahedral sp^3^ hybridization. Such conversion creates structural defects that act as additional scattering centers, thereby enhancing the D-band intensity. The appearance of new Raman bands at 980, 1148, and 580 cm^−1^, corresponding to triazine and heptazine structures [34,35], provides direct spectroscopic evidence for the incorporation of nitrogen-containing heterocycles and supports this mechanistic interpretation. According to the published data, these bands can be attributed to characteristic vibrations of triazine and heptazine structures formed as a result of thermolysis and polycondensation of melamine. The presence of these signals directly confirms the presence of nitrogen-containing heterocyclic motifs on the graphene oxide surface, which is fully consistent with the FTIR spectroscopy data where bands characteristic of C–N bonds were observed. Thus, the Raman spectroscopy results clearly indicate a structural modification of the carbon nanofiller, characterized by an increase in the imperfection of its crystal lattice and the appearance of spectral signatures of covalently bonded nitrogen-containing heterocycles, which is a key factor explaining the improved compatibility and high thermal stability of the final composite.

The identified structural changes—namely, increased imperfection of the carbon lattice and the presence of covalently bonded nitrogen-containing heterocycles on the mel-GO surface—may serve as a key factor explaining the observed differences in the reactivity and thermal behavior of the composites. These active sites are capable of catalytically influencing the curing process, initiating an earlier start and increasing its intensity due to additional interaction between the melamine amino groups and the epoxy rings of the matrix. At the same time, the modified surface creates a stronger chemical bond with the polymer, which, together with the nitrogen-carbon structures acting as thermal stabilizers and carbonization catalysts, leads to a significantly greater increase in the thermal-oxidative stability of the final composite compared to a system based on unmodified GO. While FTIR and Raman spectroscopy provide strong evidence for covalent functionalization, definitive confirmation of bonding states and quantitative determination of chemical bond types would require X-ray photoelectron spectroscopy (XPS) with high-resolution C1s and N1s spectra. Such analysis was not available in the present study due to instrumentation limitations. Nevertheless, the consistent spectroscopic, diffraction, and thermal data collectively support the conclusion that melamine is covalently attached to the GO surface through nucleophilic attack of amino groups on epoxide and carboxyl functionalities, forming C–N–C and amide linkages.

SEM data of mel-GO, Figure 11, demonstrate a typical layered-lamellar morphology with the formation of agglomerates; interlamellar voids are observed between the stacks of lamellar layers, and at higher magnification, the edges of the sheets and localized areas of delamination with thin lamellar fragments on the surface of the agglomerates are clearly visible. On the one hand, this morphology indicates the material’s continued tendency to aggregate, and on the other, the presence of zones with a more open structure and potentially increased effective interfacial area. Together with the FTIR/Raman data indicating the appearance of N-containing functional fragments after modification, these observations can be considered as structural prerequisites for a more pronounced effect of mel-OG on the epoxy system, since an increase in the proportion of the accessible surface and active interfacial centers is potentially capable of initiating polymerization processes and promoting more intense heat generation during curing, and the layered-lamellar architecture with sufficient dispersion is a more efficient implementation of the barrier mechanism under conditions of thermal-oxidative degradation, which is consistent with the increase in THRI and the shift in the TGA curve of the composite with mel-OG to the region of higher temperatures.

4. Conclusions

The study shows that functionalization of graphene oxide with melamine leads to a pronounced effect of the nanofiller on the properties of the epoxy composite compared to unmodified GO, which is consistent with changes in the structure of the filler itself and the morphology of the composite. XRD, IR and Raman spectroscopy data for mel-GO reveal signs of structural rearrangement of the carbon phase and the appearance of nitrogen-containing functional fragments, while SEM indicates a layered-lamellar morphology with localized delamination zones. For the composite, this is manifested by a change in the fracture pattern. Fractography reveals a transition from the relatively smooth cleavage surface of the pristine epoxy matrix to a more developed relief in systems with GO and especially with mel-GO. This shows an increase in the fracture energy and is consistent with the increased strength and stiffness of the composites. The GO composite slows curing, increasing the gelation time from 104 to 146 min, while mel-GO dramatically accelerates the process to 48 min and increases the maximum self-heating temperature to 122 °C. DSC data confirm this trend by shifting the reaction onset to higher temperatures for GO and to lower temperatures for mel-GO. Improved structure formation and interfacial organization can be seen in the properties. The addition of GO increases strength and tensile and flexural modulus, while the use of mel-GO provides an additional enhancement of these characteristics, accompanied by increased hardness and impact strength. According to TGA data, mel-GO significantly improves the thermal-oxidative stability of the epoxy composite, while the effect of unmodified GO remains weak. The likely reasons for the change in thermal stability include improved implementation of the barrier mechanism and increased stability of the forming coke residue under oxidative conditions due to the presence of nitrogen-containing groups on the mel-GO surface.

Thus, the proposed approach allows us to overcome the key limitations inherent in systems with unmodified graphene oxide and to purposefully design epoxy materials with a record-breaking combination of strength and thermal stabilization properties for the application under conditions of increased mechanical and thermal loads.

The promising results obtained with mel-GO/epoxy composites open several avenues for future research. Further optimization of the functionalization process for industrial scalability, achieving more uniform nanofiller dispersion, and evaluating long-term durability under service conditions would be valuable next steps. Additionally, future studies should address the limitations of the present work through: multi-heating-rate TGA for kinetic analysis of thermal degradation; XPS characterization to quantitatively confirm bonding states; DMA for crosslinking density determination; direct interfacial strength measurements; exploration of alternative nitrogen-containing modifiers; and extension to other polymer matrices.