1 Introduction

MXenes, particularly Ti₃C₂Tₓ, are two-dimensional transition metal carbides or nitrides that exhibit exceptional physical and chemical properties due to their unique layered structures. These materials are terminated with functional groups such as hydroxyl (-OH), fluorine (-F), or oxygen (-O), which influence their stability and electronic interactions. The highly ordered, sheet-like morphology of MXenes facilitates efficient phonon transport along the tightly packed layers, leading to significant in-plane thermal conductivity. Experimental studies reported an increase in effective thermal conductivity from 1.26 W·m⁻¹·K⁻¹ at 80 K to 2.84 W·m⁻¹·K⁻¹ at 290 K [1], while Ti₃C₂Tₓ thin films reached conductivities as high as 55.8 W·m⁻¹·K⁻¹ [2, 3]. Theoretical calculations suggest that specific MXenes, such as Ti₃C₂O₂, could achieve even higher thermal conductivities, up to 140.25 W·m⁻¹·K⁻¹ [4, 5]. However, the realization of these values is often hindered by structural defects introduced during synthesis [6].

In the context of high-temperature applications, SiOC ceramics are particularly attractive due to their robust covalent bonding, thermal stability, and resistance to oxidation. SiOC retains its structural integrity even in harsh oxidizing environments [7,8,9,10] and can maintain an amorphous or nanocrystalline state at elevated temperatures between 1000 °C and 1200 °C in inert or reducing atmospheres [11,12,13]. The combination of MXene’s superior thermal conductivity with SiOC’s outstanding thermal stability presents an opportunity to engineer advanced ceramic composites with enhanced thermophysical properties [14,15,16,17]. While research on MXene/SiOC composites remains limited, our recent investigation into the structural evolution and electrical conductivity of MXene/SiOC ceramics demonstrated promising potential [18]. These composites exhibit superior mechanical strength, electrical conductivity, and thermal performance, positioning them as strong candidates for high-temperature applications [18,19,20]. However, the temperature dependence of thermal conductivity in these materials is governed by a complex interplay of composition, microstructure, and phonon scattering mechanisms. Understanding these factors is crucial for optimizing heat transport in engineered ceramic systems.

To predict and optimize thermal conductivity in composite materials, various theoretical models have been employed, including the rule of mixtures [21], finite difference methods [22], effective medium theories such as Caruso’s model [23], Maxwell-Garnett theory [24, 25], and cylinder assemblage model [26]. These approaches account for key parameters such as anisotropy, porosity, and interfacial resistance, enabling accurate estimations of thermal transport behavior in heterogeneous systems.

In this study, functionalized Ti₃C₂Tₓ was incorporated into polysiloxane (PSO) in varying weight fractions (1–10 wt%), followed by pyrolysis in an Ar atmosphere at different temperatures. The bulk density of the composites was determined using Archimedes’ principle, while their chemical bonding and electronic states were characterized via X-ray photoelectron spectroscopy (XPS) and Fourier Transform Infrared Spectroscopy (FTIR). The thermal conductivity of the MXene/SiOC composites was measured using the laser flash technique up to 1000 °C. Furthermore, we discussed the theoretical models applicable to high-temperature ceramic systems, providing insights into the fundamental mechanisms governing heat transport in these novel materials.

2 Experimental

2.1 Materials

Polysiloxane (PSO, Polyramic® SPR-684, Starfire Systems, Inc., Schenectady, NY, USA) served as the precursor material for this study. A Pt catalyst (2% platinum-divinyltetramethyldisiloxane complex in xylene), procured from Gelest Inc., Morrisville, PA, was used for crosslinking PSO. Tetra-n-butylammonium hydroxide (TBAOH, [54,55,56] wt% in H2O) was purchased from Sigma-Aldrich, USA. N-(2-aminoethyl)−3-aminopropyltrimethoxysilane (AEAPTMS) and dimethyl sulfoxide (DMSO) of 99.99% purity were supplied by Fisher Scientific Company, NJ, USA.

2.2 Preparation of MXene

MXene was synthesized through etching of MAX phases using a HF/LiF/HCl solution. The specific synthesis procedure is as follows. Initially, a mixture of 10.0 g TiC, Al, and Ti in a ratio of 2:1:1 was ball-milled and then calcinated at 1380 °C under an Ar atmosphere. The resulting MAX, Ti3AlC2, was ground, sieved, and subjected to vacuum drying, and then chemically etched using a solution composed of 2 parts of LiF (2 M), 1 part of HF (1 M), and 9 parts concentrated HCl (9 M). Following etching, Ti3C2Tx nanosheets were exfoliated using freezing-thawing-sonication (F/T/S) and then functionalized with DMSO and AEAPTMS before being stored in a vacuum chamber. Our earlier work detailed the sizes and morphologies of the synthesized MXenes [27]. In this study, SEM analysis showed multi-layered MXenes, with flakes of micron-sized lateral dimensions and nanoscale thickness, which is expected since the dry sample induced stacking of individual MXene layers (Supplement, Figure S1).

2.3 Synthesis of MXene/SiOC composites

MXene/SiOC composites with 0–10 wt% MXene (in wt% relative to that of PSO) were synthesized. Singh et al. used 3D finite element simulations to show that uniformly dispersed Ti₃C₂ layers in a SiOC matrix were limited to about 0.1–1.0 wt% (relative to PSO), defining a dispersion limit rather than a percolation threshold [18]. However, literature on MXene and other 2D fillers in polymer or ceramic matrices reported that thermally (or electrically) conductive networks typically emerge at higher loadings, with effective percolation onsets in the 3–15 wt% range, where frequent filler–filler contacts enable continuous pathways. Guided by these insights, the present work explored 0–10 wt% MXene to cover both the low-loading, well-dispersed regime and the high-loading regime where network formation and matrix transformation become dominant [3, 28,29,30]. In particular, loading levels around 5–7 wt% provide strong interfacial interaction with the SiOC matrix, facilitating efficient phonon transport and enhancing overall thermal conductivity [31,32,33,34,35]. To this mixture, 50 ppm of diluted Pt catalyst (relative to PSO) was added and stirred for 3 h at 30 °C. Subsequently, the mixture was degassed in a vacuum chamber (10 mTorr) for 2 hours and cast into cylindrical disks (15 × 3 mm2). Afterwards, the mixture was placed in a vacuum for 10 min to remove residual bubbles and then kept in an oven for 12 h at 60 °C, followed by 24 h at 120 °C, and finally for 12 h at 150 °C, respectively. The cured samples were polished (P-800 SiC sand paper) and ultrasonicated in ethanol for 30 min, and then polished on both sides before their dimensions were measured.

To ensure defect-free samples after pyrolysis, each sample was covered by graphite felts and well separated during pyrolysis in a tube furnace (Rapid Temp Models 1730-12, 80 × 1000 Horizontal Tube Furnace, CM Furnaces Inc., NJ, USA) with an Ar gas flow of 900 mL⋅min− 1 up to 1500 °C. The heating rate was maintained at 1 °C⋅min− 1, and the holding time at this temperature was 2 h. The product after pyrolysis resulted in cylindrical discs approximately 8 mm in diameter and 4 mm in height.

2.4 Characterization

To characterize the functional groups on the surface of MXene and investigate the bonding interactions between MXene and SiOC, FTIR was employed. The chemical states and bonding characteristics of the MXene/SiOC composite samples were measured using X-ray Photoelectron Spectroscopy (XPS) (VersaProbe 5000, UAB, USA) with a monochromatized Al Kα X-ray source. High-resolution (HR) spectra were collected at a constant pass energy of 20 eV and quantified using empirically derived relative sensitivity factors provided by Kratos Analytical. Binding energies were referenced to the C 1 s peak of the C–C bond at 284.80 eV. The data were analyzed by the commercially available software CasaXPS. Raw fine scan spectra were not smoothened; instead, the background was subtracted and peaks were fitted using mixed Gaussian–Lorentzian line shapes. For each chemical state (C–C, C–Si, C = O, Si–C, Si–O, Ti–C, Ti–O), the binding energies and full widths at half maximum (FWHM) were restricted to literature-reported ranges for SiOC and Ti₃C₂Tₓ/TiC/TiOₓ phases [36,37,38]. Raman spectroscopy, conducted with a 532 nm excitation, was used to examine the vibrational modes of MXene and SiOC, providing insights into shifts or changes in the D and G bands of MXene, which can indicate interactions with SiOC. The surface morphology of MXene and the fractured surfaces of the MXene/SiOC composites were analyzed using a Field Emission Scanning Electron Microscope (FE-SEM, Quanta 250, USA) equipped with an energy dispersive spectrometer (EDS) at an accelerating voltage of 10 kV. Helium pycnometry was used to determine the skeletal (‘true’) density (ρ, g⋅cm−³) of the pyrolyzed ceramics, following standard practice for porous ceramics; helium gas penetrates open, connected pores, so the measured volume excludes open porosity and corresponds to the solid skeleton volume. X-ray diffraction (XRD) analysis, conducted with a CuKα radiation source (λ = 0.15406 nm) using an Empyrean XRD system (Malvern Panalytical, USA), was used to assess crystallinity changes in the functionalized MXene and MXene/SiOC samples. Thermal diffusivity (α, mm²⋅s− 1) of the MXene/SiOC composites and pure SiOC was measured with a flash thermal conductivity meter (LFA 1000, LINSEIS Inc., USA), with each sample having a size of 12 mm (diameter) × 2 mm (thickness). The thermal conductivity (λ, W·m⁻1·K⁻1) was calculated using the equation:

$$\lambda=\alpha\;\times\;p\;\times C_p$$
(1)

where α is the thermal diffusivity, ρ is the density, and Cp is the specific heat capacity. For each composition and temperature, five independent measurements of thermal diffusivity were performed, and the reported thermal conductivities corresponded to the mean values, with standard deviations used to construct the error bars. The through-plane electrical conductivity (σ) of SiOC and MXene/SiOC composites was evaluated using electrical impedance spectroscopy (EIS) with a modified four-point probe setup (VersaSTAT 3, Princeton Applied Research, Oak Ridge, TN). Direct current (DC) measurements were conducted in the potential range of ± 1.5 V at a scan rate of 0.01 V⋅s⁻¹. During the measurements, the samples were positioned between two platinum (Pt) meshes (mesh size: 0.25 mm; wire diameter: 0.06 mm; Sigma-Aldrich, Burlington, MA). Electrical contact was established using four Pt wires (8 mil, 100%; SPI Supplies, West Chester, PA) bonded with Pt paste (5544 Pt Conductor, Ferro Corporation, King of Prussia, PA). High-resolution transmission electron microscopy (HR-TEM) was also performed to investigate the fracture surface of the MXene/SiOC composites at the nanoscale for interfacial bonding.

2.5 Theoretical models for thermal conductivity

Various theoretical models have been developed to predict the thermal conductivity of polymer-based composites. However, only a few are well-suited for polymer-derived ceramic systems such as SiOC, especially when incorporating two-dimensional (2D) fillers like MXene. Among these, Nan’s model is particularly relevant for the MXene/SiOC system as it accounts for key factors like filler orientation, aspect ratio, and interfacial thermal resistance [39]. This is crucial in our case, where the mismatch in phonon spectra and bonding at the MXene/SiOC interface can significantly influence heat transport. In addition, Nielsen’s model provides insights into the role of filler geometry and packing behavior [40,41,42]. It emphasizes how high-aspect-ratio fillers such as MXene nanosheets can improve thermal pathways even below the percolation threshold by enhancing connectivity between adjacent particles. Although more empirical, Agari’s model helps describe the gradual rise in thermal conductivity with increasing filler content and is useful in characterizing the thermal behavior in the low-loading or sub-percolation regime [43,44,45].

The most notable thermal conductivity enhancement in MXene/SiOC composites arises from the formation of interconnected MXene networks, a behavior best captured by percolation theory. Due to the 2D morphology and large surface area of MXene sheets, they can effectively form thermally conductive networks within the SiOC matrix once a critical filler concentration is reached. This leads to a sharp, non-linear increase in thermal conductivity, which is a hallmark of classical percolation behavior. For MXene/SiOC systems, the percolation threshold (Vcv) is typically reported in the range of 5–13 wt%, consistent with the trends observed in other ceramic and polymer composites [39, 46, 47]. This relationship can be mathematically captured by the percolation-based thermal conductivity model, adapted from Vysotsky’s electrical percolation theory 48:

$$\:{\lambda\:}_{c}={\lambda\:}_{f}\:{\left(\frac{{{\uplambda\:}}_{p}}{{{\uplambda\:}}_{f}}\right)}^{{\left[\frac{1-\text{V}}{1-{V}_{cv}}\right]}^{n}}$$
(2)

𝜆c is the thermal conductivity of the composite, 𝜆f is the thermal conductivity of the filler (e.g., MXene), 𝜆p is the thermal conductivity at the percolation threshold, 𝑉 is the volume fraction of the filler, 𝑉cv is the critical volume fraction (percolation threshold), and 𝑛 is the percolation exponent. For a three-dimensional composite containing randomly oriented, high-aspect-ratio 2D fillers like MXene, the thermal percolation exponent is expected to fall within the range of ~ 1.6 to 2.0, approaching the universal value for 3D conductivity percolation of ~ 2.0 [49,50,51]. This range is consistent with the analyses of similar systems, such as graphene/polymer composites, where the exponent reflects the connectivity of a sheet-based network in a 3D space [52, 53].

3 Results and discussion

3.1 Phase and structural understanding

The XRD patterns in Fig. 1 reveal distinct structural evolution pathways for the pure SiOC and MXene/SiOC composites across pyrolysis temperatures (1000–1500 °C). For the pure SiOC (Fig. 1a), the absence of sharp peaks at 1000 °C and 1200 °C confirms the expected amorphous structure, while the emergence of well-defined β-SiC peaks (36.14° (111), 60° (220), 72.15° (311)) at 1500 °C reflects successful carbothermal crystallization (Eq. 3) [15, 54].

$${\mathrm{SiO}}_2\;(\mathrm{amorphous}/\mathrm{nanodomains})\;+\;3\mathrm C\rightarrow\mathrm{SiC}\;(\mathrm\beta)\;+\;2\mathrm{CO}$$
(3)
Fig. 1
Fig. 1
Full size image

(a) XRD patterns of pure SiOC and MXene/SiOC composites (5 wt% and 10 wt%) pyrolyzed at 1000 °C, 1200 °C, and 1500 °C; (b) slow scans of the SiOC and MXene/SiOC composite pyrolyzed at 1500 °C

The MXene-loaded composites exhibit two critical deviations from this behavior. First, the β-SiC diffraction peaks exhibit reduced intensity, suggesting inhibited crystallization. The MXene sheets segment the SiOC matrix into confined domains and reduce the diffusion pathways for Si and C atoms, thereby suppressing the nucleation and growth of β-SiC crystals [8, 55]. Second, the chemical activity of MXene introduces additional complexity at elevated temperatures. Titanium preferentially forms strong Ti-O bonds, which compete with the Si-O network in the SiOC matrix. This oxygen scavenging destabilizes the SiOxC4−x units that are critical for SiC formation. Concurrently, Ti can form Ti-C bonds, leading to the local formation of TiCₓ-like structures that consume carbon, which is otherwise available for β-SiC growth. These interactions result in local compositional inhomogeneities and defects, such as vacancies and interstitials, that kinetically hinder long-range ordering. These disruptions manifest as peak broadening and reduced crystallinity in the XRD profiles of the composite [56, 57]. The mismatch in thermal expansion coefficients between MXene (~ 8 × 10⁻⁶ K− 1) and the SiOC matrix (~ 3 × 10⁻⁶ K− 1) generates residual stresses, further destabilizing SiC nucleation and growth [58, 59]. At 5 wt% loading, MXene acts as a potent catalytic agent for crystallization. This is most clearly evidenced in the slow-scan data (Fig. 1b), where the 5 wt% MXene/SiOC-1500 °C composite exhibits sharper and more defined β-SiC peaks, particularly the distinct (220) and (311) doublet at ~ 47.78° and 48.76°, compared to the broad, low-intensity peaks of pure SiOC. This signifies a higher volume fraction of more crystallized β-SiC. The presence of the peak at ~ 43.39°, while attributable to β-SiC (200), also suggests the incipient formation of TiC from MXene decomposition, providing active nucleation sites that lower the energy barrier for SiC crystallization [60]. At the critical threshold of 10 wt% MXene loading, these disruptive processes overwhelm catalytic benefits. The slow-scan pattern (Fig. 1b) for the 10 wt% MXene/SiOC-1500 °C composite reveals that the already broadened SiC-related peaks also exhibit systematic shifts to higher angles (e.g., from 36.14° and 60.00° in the reference sample (pure SiOC) to 36.67°, and 61.22° for the 10 wt% MXene/SiOC-1500 °C sample). According to Bragg’s law, such peak shifts to higher 2θ correspond to a reduction in the relevant interplanar spacings \(\:{d}_{hkl}\) and thus to a compressive lattice strain \(\:\epsilon\:={\Delta\:}d/{d}_{0}\) within the SiC/SiCxOy domains. Using these peak positions and Cu Kα radiation, we estimate (see Supplementary Information, Section S2) that the relative change in \(\:d\)-spacing, and hence in the apparent lattice parameter, is on the order of − 1% to − 1.3% compressive lattice distortion. This quantitatively supports the formation of a strained (compressed) silicon oxycarbide (SiCxOy) phase and indicates lattice distortion, which we attribute to elemental interdiffusion and the incorporation of oxygen and titanium into the SiC lattice, as well as residual thermoelastic stresses and defect fields generated by the MXene-derived phases [61, 62]. The pure SiOC ceramic pyrolyzed at 1500 °C exhibits a very broad hump centered at ~ 27°. This feature is characteristic of turbostratic free carbon within the amorphous SiOC matrix, representing the (002) graphitic stacking of the carbon nanodomains. The broad, diffuse nature of this hump confirms the highly disordered and nano-crystalline state of this carbon phase. Both the 5 wt% MXene/SiOC and 10 wt% MXene/SiOC composites exhibit sharper and more well-defined peaks around 29°. The consistent shift to a higher angle and the sharpening of the feature indicate a fundamental change in the carbon structure. This peak can be attributed to the (006) reflection of the MXene-derived carbon or the formation of a more ordered graphitic carbon structure. The peak shift from 27° to 29° signifies a decreased interlayer d-spacing, indicating that titanium species from the decomposed MXene catalyze the reorganization of the native free carbon into a tighter, more crystalline graphitic structure.

Figure 2 shows the FTIR spectra of the MXene, SiOC, and MXene/SiOC composites, which are used to identify surface functional groups and assess chemical interactions between MXene and the SiOC network. Ti–O vibrations appear in the ~ 500–800 cm⁻¹ region in all MXene‑containing samples (MXene and both composites), confirming the presence of the MXene phase in the composites. A band attributable to Ti–O–Si is observed near ~ 1000 cm⁻¹ in the 10 wt% MXene/SiOC composite and is not present in the pure SiOC spectrum, supporting the formation of a chemical bridge between the MXene surface and the SiOC network at a higher loading. The main Si–O–Si stretching region of SiOC, located around ~ 1000–1150 cm⁻¹, is noticeably modified in the composites, indicating changes in the siloxane network upon incorporation of MXene and further supporting interfacial reactions. Si–H related bands near ~ 2100 cm⁻¹ are visible in SiOC and show a reduced relative intensity in the MXene/SiOC composites, suggesting partial consumption of these reactive sites during crosslinking and bonding with MXene. In addition, a broad O–H stretching band between ~ 3200 and 3700 cm⁻¹ is present in all samples and is assigned to adsorbed moisture and surface hydroxyl groups on both MXene and SiOC, consistent with residual surface terminations and physisorbed water. Taken together, the persistence of Ti–O features, the emergence of Ti–O–Si bonding, the modification of the Si–O–Si region, and the attenuation of Si–H bands provide strong spectroscopic evidence that MXene is chemically integrated into the SiOC matrix through interfacial reactions rather than behaving as an inert physical filler.

Fig. 2
Fig. 2
Full size image

FTIR spectra of functionalized MXene and crosslinked 5 wt% MXene/SiOC and 10 wt% MXene/SiOC samples

High-resolution XPS analysis reveals that MXene serves as a source of nucleation sites for SiC crystallization within the SiOC matrix, while also uncovering contrasting transformation pathways governed by MXene loading and pyrolysis temperature. Figure 3 presents deconvoluted spectra for the Si 2p, O 1 s, C 1 s, and Ti 2p peaks, illustrating the chemical evolution and interaction between MXene and the SiOC network. For the 5 wt% MXene composites, the diffraction pattern reflects controlled crystallization and effective structural integration of MXene within the SiOC matrix. At 1200 °C, the C 1 s spectrum (Fig. 3(a)) shows a distinct peak at 283.57 eV, characteristic of C–Si bonds in SiC, confirming that SiC crystallization is initiated at a temperature where pure SiOC typically remains amorphous, as evident from the XRD results (Fig. 1). A minor component at 287.9 eV, attributable to C = O bonds, is also present (intensity at ~ 689 counts), suggesting minimal surface oxidation. Upon increasing the pyrolysis temperature to 1500 °C, the intensity of the C–Si peak rises by nearly 25% (from 6000 to 7500 counts) [10, 54, 63], while the Si 2p (Fig. 3b) spectrum shows a critical shift of the lower binding energy component from 101 eV to 100.2 eV, indicating the growth and development of a more phase-pure, crystalline SiC environment. The C = O component at this temperature is slightly more pronounced (287.4 eV), which may result from the decomposition of organic residuals or very limited oxygen mobility. The complete absence of the Ti 2p signal in these samples signifies the fine dispersion and integration of MXene-derived TiC nanocrystals within the SiOC matrix, which act as effective heterogeneous nucleation sites without agglomerating. This ideal, homogeneous nanostructure facilitates efficient phonon transport [64]. However, the 10 wt% MXene/SiOC composite undergoes a distinctly different transformation, characterized by oxidative disruption and eventual phase segregation. Notably, this occurs despite the initial presence of covalent Ti–O–Si linkages, as evidenced by FTIR (Fig. 2), indicating that high-temperature degradation and agglomeration overwhelm the stability of the initial molecular interface at this critical loading. At 1000 °C, the Ti 2p spectrum (Fig. 3c) shows a pronounced Ti–C peak at 457.82 eV (intensity at ~ 1300 counts) alongside a weaker Ti–O feature near 459 eV, indicating that MXene begins to decompose into TiC while simultaneously undergoing partial oxidation. The coexistence of both peaks reveals the formation of TiC/TiOₓ domains within the SiOC matrix. Upon further heating to 1200 °C, these Ti-rich clusters grow in size, accompanied by a Si 2p signal (Fig. 3d) dominated by the Si–O component at ~ 102 eV, suggesting oxygen migration from MXene and partial disruption of the Si–O–C network. Correspondingly, the C 1 s spectrum (Fig. 3b) shows a notable decrease in the C–Si bond intensity (~ 283.5 eV, ~ 5000 counts), confirming that the oxidative environment suppresses effective SiC nucleation. Critically, a significantly more intense C = O peak emerges at 288.6 eV, providing direct chemical evidence of enhanced oxidation processes. This confirms that the oxygen released from MXene decomposition not only forms Ti–O bonds but also creates carbonyl groups within the carbon phase, corroborating the oxidative disruption of the matrix. This transformation is primarily driven by interfacial redox reactions and surface dehydration processes [65, 66]:

$${\mathrm{Ti}}_3{\mathrm C}_2{(\mathrm{OH})}_2\;\rightarrow\;{\mathrm{Ti}}_3{\mathrm C}_2\mathrm O\:+\:{\mathrm H}_2\mathrm O\uparrow\;(\mathrm{dehydroxylation})$$
(4)
$${\mathrm{Ti}}_3{\mathrm C}_2\;+\;{\mathrm{SiO}}_2\;\rightarrow\;3\mathrm{Ti}-O\:+\:2\mathrm{Si}-C\:+\:\mathrm{CO}\uparrow\;(\mathrm{interfacial}\;\mathrm{redox})\;\;\;\;\;\;\;\;\;\;$$
(5)
Fig. 3
Fig. 3Fig. 3
Full size image

High-resolution XPS spectra of the MXene/SiOC composites: (a) C 1 s and (b) Si 2p spectra of 5 wt% MXene/SiOC; (c) C 1 s, (d) Si 2p, and (e) Ti 2p spectra of 10 wt% MXene/SiOC composites after pyrolysis

These reactions signal the beginning of MXene degradation but also demonstrate interfacial chemical bonding between MXene and the SiOC matrix.

At 1500 °C, this evolution culminates in severe phase segregation. The chemical environment is reshaped by carbothermal reduction reactions. The Ti 2p spectrum shows two nearly equal-intensity peaks at 457.8 eV and 459 eV (each at ~ 1000 counts) (Fig. 3e), signifying the coexistence of TiC and Ti–O bonds, while the overall Ti intensity decreases relative to that of 1000 °C. This reduction is attributed to the formation of a graphitic carbon overlayer, evidenced by the dominant C–C peak at 284.43 eV (intensity at ~ 8000 counts, Fig. 3b), which encapsulates TiC agglomerates and weakens the Ti signal due to surface masking, a classic XPS effect. Despite this graphitization, a distinct C = O peak persists at 288.87 eV with high intensity (~ 1380 counts, Fig. 3b), indicating that the oxidation is not fully reversed and that oxygen remains chemically bound within the system, likely at the interfaces of the agglomerates. The Si 2p spectrum, with persistent components at 101–102.2.2 eV, reflects a mixed SiC/SiOₓ environment, suggesting incomplete conversion and a disrupted network [63, 67,68,69]. The persistent Ti–O signal indicates ongoing interfacial reactions [70]:

$${\mathrm{Ti}}_3{\mathrm C}_2\;+\;{\mathrm{SiO}}_2\;\rightarrow\;3\mathrm{TiO}\:+\:2\mathrm{SiC}\;\;\;\;\;\;\;\;\;$$
(6)

This reaction consumes the SiO2 phase and contributes to the formation of defective crystalline domains. These large, carbon-coated agglomerates, directly observed in SEM, act as potent phonon scattering centers.

Thus, the XPS findings reveal a composition-dependent evolution: at moderate MXene loading (5 wt%), TiC nanocrystallites act catalytically to promote homogeneous SiC crystallization and matrix densification, whereas at high MXene loading (10 wt%), excessive carbon and TiC agglomeration induce structural inhomogeneity, surface oxidation, and carbon encapsulation, ultimately compromising the composite’s thermal and structural integrity. The evolution of the C = O signal from minor in the 5 wt% MXene/SiOC composite to intense and persistent in the 10 wt% MXene/SiOC composite serves as a direct chemical tracer for the oxidative disruption that underpins this property degradation.

Figure 4 illustrates the Raman spectra of the SiOC and MXene/SiOC composites pyrolyzed from 1000 °C to 1500 °C, providing key insights into the evolution of the free carbon phase. The characteristic D band (~ 1350 cm⁻¹), signifying disordered carbon or graphitic domain edges, and the G-band (~ 1600 cm⁻¹), representing the in-plane vibration of ordered sp² carbon, are present in all samples. Their evolution reveals a clear, loading-dependent influence of MXene on the graphitization process. For the pure SiOC, the D and G bands intensify and sharpen with increasing temperature, indicating the progressive thermal ordering of the free carbon phase. This is consistent with the XRD results (Fig. 1), which show the concurrent crystallization of β-SiC at 1500 °C, suggesting a general structural ordering across both the ceramic and carbon phases.

Fig. 4
Fig. 4
Full size image

Raman spectra showing enhanced graphitic ordering of the free carbon phase in (a) SiOC, (b) 5 wt% MXene/SiOC, and (c) 10 wt% MXene/SiOC composites

The introduction of 5 wt% MXene leads to notably smooth and well-defined Raman spectra (Fig. 4(b)). The appearance of the second-order G bands at 1200 °C, earlier than that in the pure SiOC sample, and the high intensity of the D and G bands at 1200 °C provide direct evidence of enhanced and early graphitization. This is a direct consequence of the catalytic action of the finely dispersed TiC nanocrystals, whose presence was inferred from the early onset of the Si-C bond formation in the XPS results (Fig. 3(a & b)) and the sharpened SiC peaks in the XRD pattern (Fig. 1). The fact that the D and G band intensities in the 5 wt% MXene/SiOC-1500 °C sample do not surpass those of the pure SiOC-1500 °C sample aligns perfectly with the XPS exhibition of a homogeneous, well-integrated nanocomposite (Fig. 3(a & b)) where carbon graphitization and SiC crystallization occur in a balanced, confined manner, leading to an ideal nanostructure for thermal transport (see Sect. 3.3).

The 10 wt% MXene/SiOC composite exhibits a disordered carbon network (Fig. 4(c)). The D and G bands are broader, less defined, and show significant noise at 1500 °C [71, 72]. The consistent low intensity and down-shifted band positions (D-band at 1338.3 cm⁻¹, G-band at 1587.8 cm⁻¹) indicate a highly defective carbon structure. This Raman profile directly mirrors the disruptive transformation observed in XPS and XRD. The oxidative environment from excessive MXene decomposition, evidenced by strong C = O and Ti-O signals in XPS, disrupts the formation of a continuous, ordered graphitic network. Furthermore, the formation of large TiC agglomerates, confirmed by XRD (Fig. 1) and XPS (Fig. 3), creates interfaces and defects that frustrate long-range ordering in the carbon phase. The persistent noise and poor definition of the spectra at 1500 °C are characteristic of a heterogeneous system with high phonon scattering. Thus, the Raman spectroscopy results confirm that 5 wt% MXene loading optimally catalyzes carbon graphitization, while 10 wt% MXene loading introduces too much disorder. This provides a direct link between the carbon structure and the composite’s thermal properties (see Sect. 3.3), with the well-ordered carbon in the 5 wt% MXene/SiOC composite facilitating phonon transport, and the disordered carbon in the 10 wt% MXene/SiOC composite contributing to its inefficiency.

To quantify the evolution of carbon ordering, the integrated intensity ratio \(\:{I}_{D}/{I}_{G}\) was extracted from the Lorentzian fitting to the D and G bands for all samples (Table 1). For the pure SiOC sample, \(\:{I}_{D}/{I}_{G}\) increases from 2.62 at 1000 °C to 3.11 at 1200 °C and then decreases to 2.23 at 1500 °C, indicating an initial increase in disordering followed by a net ordering increase of the free carbon phase at the highest pyrolysis temperature, consistent with the sharpening of the Raman bands and the concurrent β-SiC crystallization observed in the XRD results (Fig. 1). At each temperature, the 5 wt% MXene/SiOC composites exhibit lower \(\:{I}_{D}/{I}_{G}\) values than the corresponding pure SiOC samples (e.g., 1.65 vs. 2.62 at 1000 °C and 2.06 vs. 2.23 at 1500 °C), confirming that MXene addition promotes earlier and more effective ordering of the sp² carbon network, in agreement with the smoother Raman profiles and the appearance of second-order G features in Fig. 4(b). The 10 wt% MXene/SiOC composites, however, show relatively low \(\:{I}_{D}/{I}_{G}\) values but simultaneously exhibit broad, ill-defined bands, down-shifted G-band positions, and strong spectral noise (Fig. 4(c)); these features are characteristic of a highly heterogeneous, defect-rich carbon network for which ID/IG alone underestimates the degree of structural disordering.

Table 1 Raman ID/IG ratios of SiOC and MXene/SiOC composites at different pyrolysis temperatures
Full size table

The structural ordering inferred from the Raman spectroscopy data (Fig. 4) is directly correlated with macroscopic densification, as revealed by true density measurements. The density of the MXene/SiOC composites (Table 2) exhibits a clear dependence on both filler loading and pyrolysis temperature, reflecting the underlying structural and chemical evolution. The increase in density from 1.65 g⋅cm−³ (pure SiOC) to 1.69 g⋅cm−³ (5 wt% MXene) and 1.77 g⋅cm−³ (10 wt% MXene) at 1000 °C can be attributed to the initial incorporation of dense MXene sheets into the porous, amorphous SiOC matrix, effectively filling void spaces. This is consistent with the FTIR data (Fig. 2), which confirms covalent Ti–O–Si bonding, ensuring strong interfacial integration and efficient pore-filling from the onset.

Table 2 True densities of the pure SiOC and MXene/SiOC composites
Full size table

As the temperature rises to 1200 °C, the density increases marginally across all systems, suggesting ongoing matrix densification. The most significant densification occurs at 1500 °C, where the pure SiOC reaches 1.80 g⋅cm−³, the 5 wt% MXene/SiOC composite reaches 1.99 g⋅cm−³, and the 10 wt% MXene/SiOC composite achieves the highest density at 2.09 g⋅cm−³, resulting in exceptionally dense samples. This trend is fully supported by the XRD (Fig. 1) and XPS (Fig. 3) data: the formation of dense crystalline β-SiC and TiC phases, coupled with the consumption of the amorphous oxycarbide phase and graphitization of free carbon, leads to a more efficiently packed atomic structure. The higher density of the 5 wt% MXene/SiOC composite underscores the benefit of a homogeneous distribution of nucleation sites, fostering widespread SiC crystallization without the disruptive agglomeration that, despite yielding the highest density in the 10 wt% MXene/SiOC sample, compromises its thermal property (see Sect. 3.3).

3.2 Microstructures

The pronounced increase in density with the MXene loading, particularly at 1500 °C, necessitates a microstructural investigation to understand the underlying phase distribution and integration responsible for this densification. The SEM fracture surface of the 5 wt% MXene/SiOC composite (Fig. 5(a)) exhibits a relatively homogeneous, amorphous matrix with a smooth morphology and no visible phase segregation. The corresponding EDS spectrum shows dominant Si, O, and C signals, with only a faint Ti presence (~ 1 wt%). This indicates the successful integration of MXene-derived species, which have decomposed into ultrafine TiC nanocrystals that are uniformly dispersed within the SiOC matrix. This ideal microstructure is the direct result of the covalent Ti–O–Si interfaces formed during crosslinking (FTIR, Fig. 2) and the morphological manifestation of the well-dispersed state inferred from the XPS (absent Ti signal, Fig. 3) and XRD (sharp SiC peaks, Fig. 1) results. This homogeneous nanostructure facilitates efficient phonon transport, whereas the 10 wt% MXene/SiOC composite (Fig. 5(b)) displays a highly heterogeneous fracture surface, characterized by distinct, elongated bright regions. EDS confirms these are Ti-rich agglomerates (~ 61 wt% Ti). This visual evidence confirms the phase segregation and agglomeration of the TiC/TiOₓ phases, driven by the excessive MXene loading, which exceeds the dispersion capacity of the matrix. These agglomerates are the same features identified by XRD (shifted and broadened peaks, Fig. 1), XPS (TiC/Ti–O signal and graphitic carbon encapsulation, Fig. 3), and are responsible for the disruptive phonon scattering that degrades thermal conductivity despite the higher composite density. Additional SEM micrographs and corresponding EDS elemental maps for the 5 wt% and 10 wt% MXene/SiOC composites pyrolyzed at 1200 °C are provided in the Supplementary Information (Section S3, Figure S2).

Fig. 5
Fig. 5
Full size image

SEM fracture surface and the corresponding EDS analysis of (a) 5 wt% MXene/SiOC and (b) 10 wt% MXene/SiOC composites pyrolyzed at 1500 °C. The samples are largely dense with a few closed pores

To better understand the nanoscale features responsible for the uniform fracture surface of the 5 wt% MXene/SiOC composite, TEM coupled with EDS mapping was carried out. The TEM-EDS analysis in Fig. 6 provides quantitative insight into the nanoscale structural and compositional evolution of the 5 wt% MXene/SiOC composite with increasing pyrolysis temperature. At 1200 °C (Fig. 6(a)), the microstructure exhibits wrinkled, sheet-like features that correspond to partially preserved MXene layers embedded within the amorphous SiOC matrix. The EDS elemental maps reveal a strong co-localization of Ti with Si, O, and C, indicating that the MXene surfaces are chemically integrated within the SiOC framework rather than existing as discrete particles. The EDS visually shows that Ti is concentrated within the wrinkled domains while Si and O are distributed throughout the surrounding matrix. This likely results from partial oxidation of Ti species and condensation with Si–O units during polymer-to-ceramic conversion. This strong chemical anchoring prevents phase segregation and explains the absence of distinct TiC or TiOₓ signals in XRD (Fig. 1) and the weak Ti intensity observed in XPS (Fig. 3e), indicating an atomically mixed amorphous state at this temperature [73, 74]. At 1400 °C (Fig. 6(b)), the microstructure undergoes a profound transformation into a heterogeneous nanocomposite composed of three distinct regions. The carbon-rich amorphous matrix (Area 1) shows dominant C (~ 74 wt%) with minor O (~ 18 wt%) and Si (~ 7 wt%), consistent with carbothermal reduction of Si–O bonds and free-carbon enrichment. Area 2 exhibits high Si (~ 60 wt%) and O (~ 35 wt%) contents, identifying SiOC-derived nanocrystallites, likely nano-SiC or SiO₂-rich domains, formed by phase separation of the SiOC network. Most notably, Area 3 reveals localized Ti enrichment (~ 3.5 wt%), together with Si, O, and C, signifying the presence of TiCₓ or Ti–Si–O interfacial nanophases decorating the Si-rich nanocrystals (Fig. 7). This indicates that during high-temperature pyrolysis, MXene fully decomposes, and the released Ti species diffuse and react with carbon and Si–O units, producing TiC nanocrystallites that serve as heterogeneous nucleation centers for SiC crystallization. This mechanism directly correlates with the emergence of strong SiC peaks in the XRD results (Fig. 1) at 1500 °C, confirming catalytic crystallization by MXene-derived Ti species.

Fig. 6
Fig. 6
Full size image

High resolution TEM and EDS analysis of 5 wt% MXene/SiOC composite interfaces at: (a) 1200°C and (b) 1400°C

Fig. 7
Fig. 7
Full size image

Illustration of atomic-scale evolution pathways of the MXene/SiOC composites: integrative vs. disruptive roles of MXene during pyrolysis

Quantitatively, the redistribution of Ti from layered sheets at 1200 °C to nanophase inclusions at 1400 °C signifies a complete transformation of MXene into TiC and Ti–O–Si nanodomains. These Ti-containing nanophases, spatially confined to Si-rich sites, create a finely percolated network of SiC/TiC nanocrystals embedded in a conductive carbon matrix. This configuration enhances phonon transport, reduces interfacial resistance, and increases densification (Table 2) observed experimentally for the 5 wt% MXene composites. The absence of agglomeration, compared to higher MXene loadings, ensures uniform microstructural refinement, where the nanoscale distribution of TiC and SiC phases maximizes both thermal and electrical performance, as discussed next.

3.3 Thermal and electrical properties

The thermal and electrical conductivities of the pure SiOC and MXene/SiOC composites, summarized in Fig. 8, are not governed by composition alone but by the nanoscale architecture resulting from a critical competition between conduction network integration and disruption, as visualized in Fig. 7. Beyond the room‑temperature values, the temperature dependence of the thermal conductivity (λ) provides further insight into the dominant heat transport mechanisms in these composites (Fig. 8). For all compositions, λ increases moderately with measurement temperature in the investigated range, which is consistent with a gradual reduction in the relative importance of defect and boundary scattering and a growing contribution from intrinsic phonon–phonon (normal and Umklapp) processes in the SiC/SiOC framework [75,76,77]. In high quality ceramics, Umklapp scattering typically dominates the thermal resistance at elevated temperatures, leading to a weak or saturating λ(T) behavior once the phonon mean free path becomes limited by phonon–phonon interactions rather than by microstructural features such as pores and interfaces [78]. Within this general trend, clear differences between the 5 wt% MXene/SiOC and 10 wt% MXene/SiOC composites can be discerned. The 5 wt% MXene/SiOC sample exhibits a higher absolute λ across the entire temperature range and a relatively smooth λ vs. T curve, indicating that phonon transport is governed mainly by intrinsic Umklapp scattering in a largely continuous SiC/TiC/graphitic‑carbon network, with defect and interfacial scattering playing a secondary role [79]. In contrast, the 10 wt% MXene/SiOC composite shows both a lower λ and a weaker temperature dependence, which is characteristic of systems where temperature‑independent extrinsic scattering (from pores, agglomerates, and highly defective interfaces) dominates over intrinsic phonon–phonon processes [80]. The presence of large TiC/TiOₓ agglomerates, oxygen‑rich interphases, and residual porosity, evidenced by XRD, XPS, and microscopy, therefore introduces strong interfacial and defect scattering that limits phonon mean free paths and suppresses the expected increase in λ with temperature. Pyrolysis temperature drives the carbothermal crystallization of β-SiC and the graphitization of free carbon, processes that inherently enhance phonon and electron transport by creating more ordered, periodic lattices. However, the efficiencies of these processes and thus the ultimate conductivity depend on the continuity and quality of the conductive network formed [81, 82]. Pyrolysis temperature systematically promotes carbothermal reactions, β-SiC crystallization, and graphitization of free carbon; these thermally driven structural changes enhance both phonon and electron transport when they produce continuous, low-defect pathways. This effect is most apparent for the 5 wt% MXene/SiOC composite pyrolyzed to 1500 °C, which shows the largest λ (12–13 W·m⁻1·K⁻1) and the highest σ (~ 29.9 S·m⁻1). This dual enhancement is a direct result of a continuous, three-phase percolation network. As evidenced by the XRD (Fig. 1) and TEM (Fig. 6) results, finely dispersed TiC nanocrystals, derived from the controlled decomposition of MXene, act as heterogeneous nucleation sites that catalyze a high density of well-crystallized β-SiC domains. Concurrently, XPS (Fig. 3) and Raman spectroscopy (Fig. 4) confirm the development of a highly ordered, sp²-hybridized graphitic carbon phase. This results in an integrated architecture where crystalline SiC, conductive TiC, and graphitic carbon form interpenetrating pathways, minimizing interfacial resistance and scattering for both phonons and electrons. The accompanying increase in density supports the elimination of porosity, a key factor in reducing scattering centers.

Fig. 8
Fig. 8
Full size image

High temperature thermal conductivity of the SiOC and MXene/SiOC composites pyrolyzed at: (a) 1000 °C, (b) 1200 °C, (c) 1500 °C, and (d) electrical conductivity of the samples at room temperature

However, the 10 wt% MXene/SiOC composite at 1500 °C exhibits poor conductivities (λ ~ 3–4 W·m⁻1·K⁻1, σ ~ 2.8 S·m⁻1) despite its higher TiC and carbon contents. This failure stems from network fragmentation. At intermediate temperatures (1200 °C), the excessive MXene loading leads to non-uniform MXene decomposition, forming large TiC/TiOₓ agglomerates and releasing oxygen that oxidizes the matrix, as confirmed by the strong C = O and Ti–O signals in the XPS results (Fig. 3). At 1500 °C, severe phase segregation produces large oxide‑rich agglomerates encapsulated by carbon shells, which break the continuous pathways for both thermal and electrical conduction. Although the 10 wt% MXene/SiOC composite is slightly denser and therefore less porous than the 5 wt% MXene/SiOC sample, once porosity is largely eliminated, the dominant phonon-scattering mechanisms in dense ceramics shift from pores to grain boundaries, second‑phase particles and chemically disordered interfaces; in this regime, a high density of agglomerates and heterogeneous interfaces can strongly suppress λ and σ despite the high bulk density [83, 84]. The persistence of oxygen within the carbon phase and at interfaces introduces deep-level charge traps and potential phonon scattering sites. Consequently, while individual phases like TiC are conductive, their agglomerated and encapsulated morphology prevents the formation of a continuous conductive pathway, leading to a collapse in macroscopic phonon and electron transport. This analysis clarifies that electrical conductivity serves as a highly sensitive diagnostic parameter for microstructural changes. A high σ confirms the existence of a percolating electronic network, which in these ceramics is synonymous with a high-quality phonon-transport network. The optimal 5 wt% MXene loading achieves this by balancing catalytic nucleation with homogeneous dispersion, whereas the 10 wt% loading triggers oxidative disruption and agglomeration that fragments the network. Thus, property optimization requires not just proper phase formation, but specifically a nanoscopically continuous and chemically coherent interconnected architecture.

The experimental thermal conductivity results of the MXene/SiOC composites were evaluated against a percolation-based theoretical model (Eq. 2). The model parameters, summarized in Table 3, were selected based on literature and measured composite densities. The filler volume fraction (Vf) was calculated from the measured true density of each composite using the rule of mixtures, assuming densities of 4.2 g⋅cm−³ for MXene (Ti₃C₂) below 1500 °C and 4.93 g⋅cm−³ for TiC at 1500 °C [85, 86]. A percolation threshold (Vcv) of 5 wt% (0.0203 vol/vol) and an exponent (n) of 1.8 were used, consistent with the reported values for randomly oriented 2D sheets in a three-dimensional matrix 49–51.

Table 3 Parameters for the percolation-based thermal conductivity model
Full size table

While a full quantitative decomposition of λ into individual scattering channels (Umklapp phonons, point defects, interfaces, and pores) is beyond the scope of the present work, the relative flatness of λ(T) for the 10 wt% composite suggests that extrinsic scattering (interfaces, agglomerates, pores) contributes the majority of the thermal resistance in this sample, whereas in the 5 wt% MXene/SiOC composite the dominant contribution arises from intrinsic phonon–phonon scattering in the ordered SiC/TiC/graphitic network, with defects and interfaces providing a smaller, largely temperature‑independent background. The temperature dependence of λ was interpreted within a phonon‑gas framework, where the total phonon scattering rate is written as the sum of contributions from Umklapp phonon–phonon processes, point defects, interfaces, and pores, following standard Callaway‑type models [76, 77].

The calculated theoretical thermal conductivities are presented in Table 4. The percolation model, employing a threshold of 5 wt% MXene (Vcv ~0.0203) and standard parameters (MXene λf ~ 55 W·m⁻¹·K⁻¹, exponent n ~ 1.8), predicts a modest increase in composite conductivity from approximately 2.11 W·m⁻¹·K⁻¹ at 5 wt% MXene loading to 2.40 W·m⁻¹·K⁻¹ at 10 wt% MXene loading, based solely on the increasing volume fraction of the conductive ‘filler’ above the percolation threshold. The experimental data, where the 5 wt% MXene/SiOC composite achieves a conductivity of ~ 12.0 W·m⁻¹·K⁻¹ (~ 5.7 times of the predicted value), and the 10 wt% MXene/SiOC composite drops to ~ 3.5 W·m⁻¹·K⁻¹, clearly demonstrate that the phonon transport in this system is not governed by classical filler percolation. This significant deviation provides quantitative proof of a more complex thermal conductivity mechanism. The results highlight the dual chemical role of MXene-derived species. At the optimal 5 wt% MXene loading, MXene decomposes into catalytically active TiC nanocrystals that fundamentally transform the insulating SiOC matrix into a conductive nanocomposite by enhancing β-SiC crystallization and carbon graphitization, as proven by XRD (Fig. 1), XPS (Fig. 3), and Raman spectroscopy (Fig. 4). This creates a new, highly conductive hybrid matrix with properties far exceeding the predictions based on filler addition alone. At 10 wt% MXene loading, however, the excessive filler concentration triggers oxidative disruption and severe agglomeration, as evidenced by XPS (Fig. 3) and SEM (Fig. 5). This fragments the conductive network and introduces interfacial scattering sites, degrading the thermal conductivity despite the higher filler content. The systematic deviation from the percolation model at both loadings provides compelling quantitative evidence that the thermal properties are predominantly controlled by MXene-induced chemical modifications and the microstructural evolution of the matrix, establishing a paradigm where ‘filler’ functionality transcends its role as a mere conductive additive.

Table 4 Theoretical thermal conductivities calculated by the percolation model
Full size table

4 Conclusions

This work demonstrates that MXene loading dictates the structural evolution and functional properties of polymer-derived SiOC ceramics through a critical transition from microstructural integration to oxidative disruption and phase segregation. MXene sheets form covalent Ti–O–Si interfaces with the preceramic polymer network, which anchors their dispersion. Upon pyrolysis, this configuration enables MXene to decompose into well-dispersed TiC nanocrystals, which act as potent nucleation sites, catalyzing the premature crystallization of β-SiC at 1200 °C and enhancing the graphitic ordering of the free carbon phase. The resulting homogeneous nanocomposite forms an integrated, three-phase architecture comprising crystalline β-SiC, conductive TiC, and percolating graphitic carbon with 5 wt% MXene loading. This ideal nanostructure yields a synergistic 10% increase in density and, at 1500 °C, exceptional multifunctional transport properties: a thermal conductivity of ~ 12.0 W·m⁻¹·K⁻¹ and an electrical conductivity of ~ 30 S·m⁻¹. This performance surpasses the predictions of classical percolation models by a factor of ~ 5.7, providing quantitative evidence that property enhancement is driven not by a conductive filler network, but by MXene-induced chemical transformation of the matrix itself. However, the 10 wt% MXene loading exceeds the dispersion capacity, leading to oxidative disruption of the SiOC network, severe TiC/TiOₓ agglomeration, and the formation of defective, oxygen-rich interfacial phases. These large agglomerates fragment the conductive percolation network, which is proven by the concurrent decreases in thermal conductivity (~ 3.5 W·m⁻¹·K⁻¹) and electrical conductivity (~ 2.8 S·m⁻¹). This highlights that property enhancement is not governed by filler content alone, but by the ability to achieve a nanoscopically continuous, chemically coherent, and crystallographically ordered composite architecture. The synergy achieved at 5 wt% MXene loading establishes a new SiOC composite design principle, transforming the 2D filler from a passive additive into an active architect for synthesizing high performance, multifunctional ceramic nanocomposites.

Our results therefore identify an intermediate MXene content of 5 wt% as an optimal design point that balances catalytic crystallization and structural integrity, providing a general “sweet-spot” guideline for tailoring MXene loading in future multifunctional ceramic composites beyond the specific Ti₃C₂Tₓ/SiOC system studied here. Nevertheless, this study is limited to a single MXene chemistry (Ti₃C₂Tₓ), one polymer-derived ceramic matrix (SiOC), and a specific pyrolysis route and temperature window; the optimal loading window and structure–property relationships may shift for other MXene compositions, surface terminations, ceramic matrices, and processing conditions. Future work should therefore extend this intermediate-loading design concept to different MXene/ceramic combinations and evaluate additional functionalities (e.g., dielectric response, EMI shielding, mechanical reliability and oxidation resistance) in order to build broadly applicable guidelines for MXene-engineered multifunctional ceramic composites.