Structure–Property Relationships in Cyanate Ester Composites Incorporating BaTiO3 and Transparent Glass Fillers

TL;DR

This study explores how adding BaTiO3 and transparent glass fillers to cyanate ester composites can improve their dielectric properties for high-frequency electronics.

Contribution

The novel use of transparent glass and BaTiO3 fillers with interfacial engineering to precisely tune dielectric properties in cyanate ester composites.

Findings

TESPI surface modification improves interfacial bonding and filler dispersion in the composite matrix.

A 15 wt% filler loading achieves optimal dielectric performance with a low loss of tan δ ≈ 0.0047.

Excessive filler loading causes agglomeration and increased dielectric loss.

Abstract

Polymer–ceramic composites based on cyanate ester resins have attracted increasing attention for high-frequency electronic applications due to their low dielectric loss, thermal stability, and dimensional reliability; however, achieving a targeted dielectric constant while maintaining low loss remains a key challenge. In this study, transparent glass powders and BaTiO3 ceramic fillers were incorporated into a cyanate ester matrix to systematically investigate structure–property relationships and optimize dielectric performance for antenna-related applications. Transparent glass powders were synthesized via a melt-quenching route and combined with submicron BaTiO3 particles, while both fillers were surface-modified using 3-triethoxysilylpropyl isocyanate (TESPI) to enhance interfacial compatibility. Composite samples containing 5–30 wt% total filler were fabricated and characterized by XRD, FTIR, tensile testing, dielectric measurements, and SEM/EDX analyses. The results demonstrate that TESPI surface modification promotes strong interfacial bonding and homogeneous filler dispersion within the cyanate ester matrix. An optimal balance between mechanical integrity and dielectric performance was achieved at 15 wt% total filler loading (K3), exhibiting a dielectric constant close to 10 and the lowest dielectric loss (tan δ ≈ 0.0047 at 1 MHz). Microstructural observations confirm that excessive filler loading leads to agglomeration and increased dielectric loss. Overall, the combined use of transparent glass and BaTiO3 fillers, together with effective interfacial engineering, enables precise tuning of dielectric properties in cyanate ester composites for high-frequency electronic applications.

Figures

Full text

1. Introduction

Polymers generally exhibit good flexibility and processability; however, they typically possess relatively low dielectric constants. In contrast, ceramic materials are characterized by high dielectric permittivity, but they are inherently brittle and require high-temperature sintering for densification. As a result, the standalone use of either polymers or ceramics is significantly limited in practical applications. Polymer–ceramic composites, which integrate the complementary advantages of both constituents, have therefore emerged as promising materials with enhanced functional performance [1,2,3,4,5,6,7]. Cyanate ester-based resins have been extensively investigated in recent years because of their remarkable thermal stability, inherently low dielectric losses, elevated glass transition temperatures (Tg), and superior resistance to dimensional changes [8,9,10,11,12,13,14]. With their low dielectric loss and reduced moisture uptake, cyanate ester resins demonstrate stable performance over a broad range of temperatures and electromagnetic frequencies. For this reason, they are commonly utilized in electronic devices and systems [15,16,17,18]. One of these important application areas is GNSS (Global Navigation Satellite System) antenna systems, where material stability, low dielectric loss, and dimensional reliability are critical to ensure signal integrity, phase stability, and positioning accuracy under varying environmental conditions. For antenna applications, the dielectric constant is targeted to be in the range of 8.5–9.5, while the dielectric loss tangent is aimed to be within 0.001–0.0001 [16,17,18].

In this study, a hybrid composite system consisting of cyanate ester resin (matrix), transparent borosilicate glass powder (low-loss dielectric filler), BaTiO_3_ (high-permittivity ceramic filler), and TESPI silane coupling agent (interfacial modifier) was designed and investigated. Transparent glass powders were incorporated to improve dielectric performance and interfacial compatibility, while BaTiO_3_ was introduced to explore the synergistic effects of hybrid fillers in achieving the targeted dielectric constant. The transparent glass phase was chosen because of its relatively low dielectric constant (5.0–6.0) and very low dielectric loss (0.0008 at 1 MHz), enabling precise tuning of the dielectric behavior and suppression of the loss tangent without compromising the uniformity of the polymer matrix. BaTiO_3_, with high permittivity that strongly depends on particle/grain size, crystal structure, and frequency, was used as a high dielectric constant filler; for submicron (~500 nm) BaTiO_3_, room-temperature permittivity is typically reported in the few-hundreds range [19]. To further improve interfacial compatibility, silane coupling agents were employed to strengthen the interfacial interactions between the cyanate ester matrix and the fillers [20]. A bifunctional silane, 3-triethoxysilylpropyl isocyanate (TESPI), was used as both a crosslinking agent and an interfacial modifier and was specifically applied to surface-modify the glass and BaTiO_3_ powders. Upon hydrolysis of the ethoxy groups, TESPI forms silanol species that condense with hydroxyl groups on the filler surfaces, while the isocyanate functionality participates in chemical interactions with the cyanate ester matrix, thereby establishing covalent interfacial bridges.

Cyanate ester resins are high-performance thermoset polymers that provide dielectric properties particularly suitable for electronic applications. Cyanate ester matrices exhibit very low dielectric loss (tan δ) and low dielectric constant values, which constitute a critical advantage for materials operating at high frequencies with minimal energy dissipation and signal distortion. In addition, cyanate esters possess high thermal stability, low moisture uptake, and excellent dimensional reliability, making them well suited for microelectronic and high-frequency circuit applications. For these reasons, cyanate ester resins demonstrate superior dielectric performance compared to other thermoset resins and are therefore widely preferred in electronic composite applications. These characteristics represent the primary physical justification for selecting cyanate ester resins as the matrix material in high-frequency dielectric composites: low loss and good frequency stability minimize signal degradation and enhance device efficiency [7,8]. BaTiO_3_ ceramics are well known as perovskite-structured ferroelectric materials with extremely high dielectric permittivity, which makes them an ideal filler phase for composites designed to enhance dielectric performance [21,22]. Since the dielectric constant of the ceramic phase is significantly higher than that of the polymer matrix, the incorporation of BaTiO_3_ particles can markedly increase the overall dielectric constant of the composite. Furthermore, the microstructure and dispersion of BaTiO_3_ strongly influence polarization behavior within the composite; in particular, surface modification or appropriate processing methods that improve particle connectivity can lead to a substantial enhancement of dielectric permittivity. This phenomenon can be attributed to the more efficient contribution of the ceramic phase to the electric field. Consequently, the use of BaTiO_3_ in polymer–ceramic composites offer a highly desirable performance profile for high-frequency electronic components—such as antenna substrates and integrated circuit dielectrics—by enabling a balance between high dielectric constant and low dielectric loss.

2. Materials and Methods

2.1. Materials

The experimental materials used for preparing the reinforced composite samples in this study included cyanate ester resin, glass and BaTiO_3_ powders. A two-part cyanate ester–epoxy resin system (viscosity: 140 cP at 43 °C; cured density: 1.25 g/cm^3^) supplied by Lepus Chemical (Tekirdağ, Turkey) was employed. TESPI (Merck, Darmstadt, Germany) (Figure 1) was used as a silane coupling agent. BYK-066 N (BYK Additives, Wesel, Germany) was incorporated as a defoaming agent at 1.0 wt%. Ethanol (99% purity; Lepus Chemical, Turkey) was used as the solvent in all formulations. BaTiO_3_ was provided from a chemical company in Turkey (Lepus Chemical). The average diameter of BaTiO_3_ is 500 nm.

2.2. Transparent Glass Powder Preparation

The raw materials were weighed according to the components and specified quantities in Table 1. All raw materials were obtained with 99.9% purity (Lepus Chemical, Turkey).

After weighing, the samples were dry-mixed and subjected to melting at 1600 °C for 2 h (Nannetti Furnace, Faenza, Italy) using alumina crucibles. Then, the molten material was quenched in water to form frit. The fritted glass composition was subsequently ground for 40 h (Nanomultimix, 50S Model, Istanbul, Turkey).

2.3. Specimen Preparation

Cyanate ester resin (A/B = 2:1, w/w) was used as the polymer matrix. Cyanate ester component Resin A is a low-viscosity resin used to improve processability, whereas Resin B is a higher-functionality cyanate ester designed to enhance crosslink density and thermal stability. To enhance filler–matrix interactions, a silane coupling agent, 3-triethoxysilylpropyl isocyanate (TESPI, 1–3 wt%), was employed as both a crosslinking agent and an interfacial modifier. Prior to composite fabrication, glass and BaTiO_3_ powders (5–20 wt%) were dispersed in ethanol and surface-treated with TESPI. After solvent evaporation, the modified fillers were dried and subsequently incorporated into the cyanate ester resin.

The resin was preheated to 90 °C and stirred for 1 h at 800 rpm using an IKA RCT Basic stirrer (IKA Turkey, Istanbul, Turkey) to obtain a homogeneous liquid phase. Modified powders were then added to the resin matrix at loadings of 5, 10, 15, and 20 wt%. Subsequently, TESPI was slowly added and the mixture was stirred for an additional hour to promote interfacial reactions. Finally, the second component (B) of the cyanate ester resin was incorporated, followed by the addition of a defoaming agent (BYK-066N, BYK Additives, Wesel, Germany). The detailed compositions of all samples are summarized in Table 2. Cyanate ester components Resin A and B were used in a constant ratio throughout all compositions to maintain consistent matrix chemistry, while the BaTiO_3_ content was systematically varied. The compositions were calculated on a weight-percentage basis with respect to the total composite mass. While the glass content was kept constant, the BaTiO_3_ loading was adjusted to obtain total filler contents in the range of 5–30 wt%.

After 15 min of mixing, the resulting suspensions were cast into high-temperature-resistant silicone molds (4 cm × 8 cm × 0.5 cm) and subsequently degassed under a vacuum of approximately 25 mmHg at 60 °C for 1 h using a Weightlab WF-HTV25 oven (Weightlab, Istanbul, Turkey). The molds were then transferred to a preheated oven maintained at 60 °C. Subsequently, the samples were cured according to the following schedule: 60 °C for ½ h, 120 °C for ½ h, 150 °C for 1 h, and 180 °C for 3 h. After curing, the samples were cooled to room temperature at a controlled rate of 5 °C/min. The cured specimens were sectioned into the desired dimensions using a diamond-blade cutting system (TechCut 5™ Precision Sectioning Machine equipped with a low-concentration metal-bonded diamond wafering blade, Allied High Tech. Products, Inc., Rancho Dominguez, CA, USA). Prior to characterization, all specimens were vacuum-dried at 80 °C for 6 h and stored under dry conditions.

2.4. Characterization

Samples were characterized using a combination of structural, spectroscopic, mechanical, electrical, and microstructural analyses. Phase identification was performed by X-ray diffraction (XRD) using a PANalytical Empyrean diffractometer (Malvern Inc., Malvern, UK) operated at 40 kV and 40 mA with Cu Kα radiation (λ = 0.154 nm), collecting data over a 2θ range of 5–70° at a scanning rate of 2°/min. Fourier transform infrared spectroscopy in attenuated total reflectance mode (FTIR-ATR) was conducted using a PerkinElmer Spectrum 100 spectrophotometer (PerkinElmer, Waltham, MA, USA) in the wavenumber range of 550–4000 cm^−1^ with a spectral resolution of 4 cm^−1^, averaging 32 scans per measurement. Tensile properties were evaluated using an MTS 810 universal testing machine in accordance with GB 1040.1-2025 [23], employing dumbbell-shaped specimens tested at a constant crosshead speed of 2 mm/min at 25 °C, with reported values representing the average of five samples. The relative dielectric permittivity (ε_r_) and dielectric loss (tan δ) were measured under ambient conditions using a precision LCR meter (Hioki3532-50, Hioki E.E. Corporation, Nagano, Japan) over a frequency range of 100 kHz to 1 MHz; prior to testing, the sintered disk samples were polished and coated with silver paste on both surfaces to form electrodes. For each composition, dielectric measurements were performed on at least three independently prepared specimens, and the reported values represent the average of these measurements. Finally, the microstructural features and phase distribution of the cured samples were examined by scanning electron microscopy (SEM) coupled with energy-dispersive X-ray spectroscopy (EDX) using a Philips XL30 SFEG scanning electron microscope (Philips, Hillsboro, OR, USA).

3. Results and Discussion

3.1. XRD Analysis Results of the Specimens

The X-ray diffraction (XRD) pattern of the synthesized T-glass powder (Figure 2) exhibits a wide diffuse halo in the 25–30° (2θ) region, while no well-defined crystalline reflections are detected over the measured angular range of 10–70°. Such a diffuse scattering feature is indicative of the amorphous nature of the material [24,25].

The XRD analysis of the commercially available BaTiO_3_ powder was also performed, and the corresponding diffraction pattern is presented in Figure 3. The obtained diffraction pattern and inter-planar spacings match well with those reported for BaTiO_3_ in the JCPDS database (card No. 75–1606) [26]. A discernible peak splitting/shoulder observed in the 40–50° (2θ) region, particularly around ~45°, is characteristic of tetragonal BaTiO_3_ and can be indexed to the (002)/(200) reflections. In contrast to cubic BaTiO_3_, which exhibits a single symmetric (200) reflection in this region, the observed splitting indicates the predominance of the tetragonal perovskite structure in the BaTiO_3_ powder [27].

Figure 4 shows the stacked XRD patterns of the standard cured cyanate ester matrix (STD) and the composite samples containing BaTiO_3_ and transparent glass fillers (K1–K6). The diffraction pattern of the STD sample exhibits a broad diffuse halo centered in the 8–20° (2θ) range, which is typical of an amorphous polymeric structure. The absence of sharp diffraction peaks confirms that the standard cured cyanate ester is fully amorphous and does not contain any crystalline phases. In contrast, all composite samples (K1–K6), which incorporate BaTiO_3_ and transparent glass fillers, display well-defined and sharp diffraction peaks superimposed on the amorphous background. The crystalline reflections, highlighted by star symbols in the figure, can be indexed to tetragonal BaTiO_3_. A notable feature is the peak splitting or shoulder observed near 45° 2θ, attributed to the overlapping (002) and (200) reflections. This characteristic splitting is a definitive indicator of the tetragonal perovskite structure of BaTiO_3_ and confirms that the crystalline phase is preserved after incorporation into the cyanate ester–glass matrix. The transparent glass filler does not introduce additional crystalline peaks, indicating that it remains predominantly amorphous within the composite structure. As the filler content increases from K1 to K6, the intensity of the BaTiO_3_ diffraction peaks increases progressively, while their positions remain unchanged. This observation suggests an increased contribution of the crystalline BaTiO_3_ phase without any phase transformation, secondary phase formation, or structural degradation during the composite preparation process. The coexistence of the broad amorphous halo from the cyanate ester and glass phases with the sharp BaTiO_3_ reflections demonstrates the successful formation of a hybrid amorphous–crystalline composite system.

3.2. FTIR Analysis Results of the Specimens

In polymer matrix composites containing ceramic fillers, achieving homogeneous dispersion of high-density ceramic particles within the polymer matrix and preventing particle sedimentation and agglomeration remain major challenges. To overcome these issues, glass and BaTiO_3_ powders were surface-modified with TESPI in this study. The effect of TESPI treatment on BaTiO_3_ ceramic particles was also investigated. Figure 5 compares the FTIR spectra of unmodified BaTiO_3_ and TESPI-modified BaTiO_3_ (BaTiO_3_ + TESPI). Although the overall spectral profiles remain similar, several noticeable differences are observed after TESPI modification, indicating changes in the surface chemistry of BaTiO_3_. In particular, variations in the 1000–1200 cm^−1^ region are attributed to the formation of Si–O–Ti and Si–O–Ba bonds, suggesting silane coupling between TESPI and the BaTiO_3_ surface. Additionally, changes observed around ~1700–1730 cm^−1^ are associated with the formation of urethane-type linkages resulting from the reaction between the isocyanate groups of TESPI and surface hydroxyl groups on BaTiO_3_. In the modified BaTiO_3_ sample, slight spectral variations are observed in the 2100–2200 cm^−1^ region compared to the unmodified powder. However, because a weak band is also present in the pristine BaTiO_3_ sample, this region alone cannot be considered definitive evidence of surface grafting. Owing to the relatively low grafting density and overlapping vibrational bands, the spectral differences remain subtle. Nevertheless, the combined variations observed in the 1000–1200 cm^−1^, ~1700–1730 cm^−1^, and 2850–2950 cm^−1^ regions collectively indicate surface modification of BaTiO_3_ with TESPI. The band at ~2850–2950 cm^−1^ corresponds to aliphatic C–H stretching vibrations of TESPI, indicating the presence of organic silane groups on the BaTiO_3_ surface [28,29,30,31]. These findings, together with literature reports on silane-treated ceramic fillers [29,30,31,32], suggest improved interfacial compatibility between the modified BaTiO_3_ particles and the polymer matrix.

After the modification of the ceramic powders, FTIR analyses of the cured samples were carried out and compared with the standard (STD) sample, as shown in Figure 6. FTIR spectra of the cured samples were analyzed to evaluate the effect of TESPI-modified ceramic fillers on the chemical structure and interfacial interactions within the cyanate ester matrix. The spectrum of the standard sample (STD), which contains no inorganic fillers, is primarily characterized by the typical absorption bands of the cured cyanate ester network, indicating the formation of a crosslinked polymer structure without contributions from inorganic phases. In contrast, the composite samples (K1–K6), incorporating increasing amounts of TESPI-modified glass and BaTiO_3_ fillers, exhibit noticeable changes in their FTIR spectral features. In particular, the progressive enhancement of absorption bands in the 1000–1150 cm^−1^ region, attributed to Si–O–Si and Si–O–M (M = Ba, Ti) stretching vibrations, confirms the presence of ceramic fillers and their effective chemical integration into the polymer matrix. Additionally, variations observed around ~1700–1730 cm^−1^ are associated with urethane-type linkages formed through reactions between TESPI-derived isocyanate groups and surface hydroxyl groups on the fillers [31,32,33], indicating the establishment of covalent interfacial bonding. The absence of significant reduction in the characteristic –NCO stretching band near ~2270 cm^−1^ in all composite samples further confirms the consumption of free isocyanate groups during surface modification and composite formation [34,35,36,37,38,39]. As the total filler content increases from K1 to K6, the intensity of inorganic-related bands becomes more pronounced, reflecting the higher contribution of the ceramic phases, while the retention of the polymer backbone features indicates successful curing of the cyanate ester matrix in all formulations. Overall, the FTIR results demonstrate that TESPI-modified ceramic fillers are chemically grafted within the cyanate ester matrix, leading to the formation of a well-defined organic–inorganic hybrid structure across all composite samples.

3.3. Tensile Strength Results

Figure 7 presents the tensile stress–strain (σ–ε) curves of the standard cyanate ester matrix (STD) and the K1–K6 composite samples containing different proportions of glass powder and BaTiO_3_. The graph comparatively illustrates the effect of increasing filler content on the maximum tensile stress and elongation at break of the composites. For all samples, the stress exhibits an approximately linear increase in the low-strain region; however, depending on the filler content, noticeable variations are observed in the slope of the curves, the maximum stress values, and the fracture behavior. Examination of the graph reveals that the neat cyanate ester matrix (STD) exhibits the lowest elastic modulus and maximum tensile stress, while displaying the highest elongation at break. This behavior indicates a more ductile response to the unfilled matrix, which can be attributed to the high mobility of polymer chains. In the K1 and K2 samples (5–10 wt% total filler), a pronounced increase in elastic modulus and maximum tensile stress is observed with the incorporation of BaTiO_3_ and glass powder, accompanied by a gradual decrease in elongation at break. This trend suggests that the filler particles begin to effectively transfer load to the matrix, while the increasing ceramic content progressively restricts polymer chain mobility. The K3 sample (15 wt% total filler) exhibits the most balanced mechanical performance, combining a high maximum tensile stress with a relatively preserved elongation at break. This result indicates that the filler–matrix ratio in this composition is at an optimal level and that effective interfacial bonding is achieved through the homogeneous dispersion of BaTiO_3_ and glass powder particles within the matrix, facilitated by TESPI surface modification. In contrast, although the maximum tensile stress continues to increase in the K4–K6 samples (20–30 wt% total filler), a pronounced reduction in elongation at break and a more abrupt fracture behavior are observed. This behavior can be attributed to increased filler–filler interactions and possible particle agglomeration at high filler loadings, leading to stress concentration and a more brittle composite response. Overall, the most favorable strength–ductility balance from a mechanical perspective is achieved in the K3 composite.

3.4. Dielectric and Tangent Loss Evaluation

Figure 8 presents the variation in the dielectric constant (ε_r_) of the standard cyanate ester (STD) and the developed composite formulations (K1–K6) as a function of frequency in the range of 200 kHz to 1 MHz. The neat cyanate ester (STD) exhibits a relatively low dielectric constant of approximately 2.7–2.9 over the entire frequency range, confirming its intrinsically low permittivity and weak polarization capability. In contrast, all composite samples show a significantly higher dielectric constant, primarily due to the incorporation of high-permittivity BaTiO_3_ particles together with glass powder. Among the composites, K3 and K4 display the highest dielectric constants, reaching values close to 10 at lower frequencies, which can be attributed to their higher ceramic loading and more effective interfacial polarization between the polymer matrix and ceramic fillers. The increase in εᵣ with filler content indicates that the dielectric response is governed by Maxwell–Wagner–Sillars (MWS) interfacial polarization, arising from charge accumulation at the polymer–ceramic interfaces [40,41,42,43].

Figure 9 illustrates the frequency-dependent dielectric loss (tan δ) behavior of the STD and K-series samples (K1–K6) over the frequency range up to 1 MHz. At low frequencies, all samples exhibit relatively high and fluctuating loss values, which can be attributed to interfacial polarization, space-charge effects, and dipolar relaxation mechanisms that dominate in this region. As the frequency increases, the loss tangent of all compositions decreases significantly and gradually approaches a stable plateau, indicating the suppression of interfacial polarization and reduced dipole mobility at higher frequencies.

For each composition, dielectric measurements were performed on at least three independently prepared specimens, and the reported values represent the mean of these measurements. The experimental variation between samples was found to be small, confirming the reproducibility of the results and supporting the identification of the optimal composition (K3). Among all compositions, K3 exhibits the lowest dielectric loss across the investigated frequency range, demonstrating superior dielectric stability compared with the other samples. In particular, at 1 MHz, the tan δ value of K3 is approximately 0.0047 ± 0.0010, which is markedly lower than those of the remaining K-series samples and the STD reference.

The dielectric loss trends presented in Figure 10 are closely associated with the compositional differences among the samples. At lower filler loadings (K1 and K2; 5–10 wt%), only a minor decrease in tan δ is observed relative to the STD sample, suggesting that the BaTiO_3_ concentration is not sufficient to markedly influence polarization mechanisms or effectively limit molecular motion within the polymer matrix. By contrast, K3, containing 15 wt% total filler, emerges as a threshold composition at which an effective equilibrium between the polymer phase and ceramic inclusions is established. At this filler level, the BaTiO_3_ particles are present in adequate quantity to hinder dipolar relaxation and suppress interfacial charge buildup, while maintaining a uniform dispersion due to the synergistic effect of TESPI-induced surface modification. As a result, interfacial polarization losses are minimized and interfacial adhesion is improved, yielding the lowest dielectric loss among all investigated compositions. Accordingly, K3 demonstrates a tan δ value of approximately 0.0047 ± 0.0010 at 1 MHz, indicating excellent dielectric stability under high-frequency electric fields. When the filler loading surpasses 15 wt% (K4–K6; 20–30 wt%), the dielectric loss shows a progressive increase. This behavior is primarily attributed to intensified filler–filler interactions, partial agglomeration of BaTiO_3_ nanoparticles, and the diminished continuity of the polymer matrix. Such microstructural changes enhance interfacial heterogeneity and facilitate localized charge trapping, leading to increased energy dissipation, particularly in the low- and mid-frequency regions. To better position the present study within the literature, the dielectric performance obtained here was compared with previously reported cyanate ester-based and ceramic-filled polymer composites [2,7,19]. The composites developed in this study exhibit a balanced combination of moderate permittivity and low dielectric loss, particularly for the K3 composition at 1 MHz, which is comparable to recent low-loss cyanate ester composite systems [13,14]. These results demonstrate the effectiveness of the proposed formulation and support the novelty of the study.

3.5. SEM/EDX Results

Figure 10 presents the SEM micrograph of the surface morphology of the standard cyanate ester resin after curing. The cured surface exhibits a relatively dense and continuous microstructure without visible macro-scale defects such as cracks or voids, indicating effective curing and network formation.

Figure 11 presents the surface morphology of the cured K-series composites (Figure 11a–f). At lower filler loadings (K1 and K2; Figure 11a,b), the surfaces appear relatively smooth and continuous; however, shallow grooves and localized heterogeneities are still visible, indicating a less compact surface structure. In contrast, K3 (Figure 11c) exhibits a more compact and uniform surface texture compared with K1 and K2. This observation suggests improved matrix–filler interaction and a reduced degree of surface heterogeneity at this composition. With further increase in filler loading (K5 and K6; Figure 11e,f), larger surface clusters and irregular features become more apparent.

Figure 12 shows the cross-sectional SEM images of the developed K1–K6 (Figure 12a–f) compositions. The micrographs reveal the distribution of the filler phases within the cyanate ester matrix. The dashed rectangles indicate regions where filler particles are locally concentrated, suggesting the formation of agglomerated or particle-rich zones in the section area. In K1 (Figure 12a) and K2 (Figure 12b) samples, relatively large particle clusters are observed, indicating non-uniform filler distribution. The K3 composition (Figure 12c) exhibits a more homogeneous microstructure, where the fillers appear to be more evenly dispersed within the matrix. In K4 (Figure 12d) and K5 (Figure 12e), multiple particle-rich regions can be seen, showing partial agglomeration of the filler phase. Similarly, K6 (Figure 12f) also presents localized particle accumulation, although the matrix–filler interface remains continuous. The agglomeration of ceramic fillers observed in the SEM micrographs is consistent with literature reports indicating that strong interparticle interactions and high surface energy promote clustering in highly filled polymer systems [44,45,46,47,48]. Although silane-based surface modification improves interfacial adhesion, complete elimination of agglomeration is rarely achieved in ceramic-filled composites [45,46,47,48]. Overall, the cross-sectional microstructural observations strongly support the dielectric and tan δ results, confirming that K3 (Figure 12c) represents the optimal composition. The balanced filler loading in K3 ensures uniform dispersion, strong interfacial adhesion, and a continuous polymer network, which collectively explains its superior dielectric performance and lowest loss tangent among all investigated samples.

Figure 13a shows the SEM image and corresponding EDX analyses (Figure 13b,c) of the best-performing composite sample (K3), which contain BaTiO_3_ particles and transparent glass fillers in the cyanate ester matrix. The SEM micrograph (Figure 13a) reveals a dense and homogeneous microstructure, and two representative regions (marked as 1 and 2) are selected. The EDX spectrum obtained from region 1 (Figure 13b) is characterized by intense barium (Ba) and titanium (Ti) peaks, confirming the presence of BaTiO_3_ particles within the composite structure. The clear and well-defined peaks indicate that the BaTiO_3_ phase remains chemically stable and well preserved after processing and curing. The EDX spectrum collected from region 2 (Figure 13c) shows dominant carbon (C) and silicon (Si) signals, accompanied by minor contributions from aluminum (Al) and calcium (Ca). The strong carbon peak is attributed to the cyanate ester polymer matrix, while the pronounced silicon signal originates from the transparent glass phase as well as the silane (TESPI) coupling agent used for surface modification. The presence of Al and Ca further supports the incorporation of the glass filler within the polymer matrix. The EDX analysis demonstrates the successful coexistence and homogeneous distribution of BaTiO_3_ particles and glass fillers in the cyanate ester matrix.

4. Conclusions

In this study, cyanate ester-based polymer–ceramic composites incorporating transparent glass and BaTiO_3_ fillers were successfully developed, and their structure–property relationships were systematically evaluated. Surface modification of both fillers with 3-triethoxysilylpropyl isocyanate (TESPI) proved to be an effective strategy to enhance interfacial adhesion and promote homogeneous dispersion within the polymer matrix, as confirmed by FTIR and SEM/EDX analyses. The incorporation of transparent glass enabled controlled tuning of the dielectric response while suppressing dielectric loss, whereas BaTiO_3_ provided the necessary enhancement in dielectric permittivity. Among the investigated formulations, the composite containing 15 wt% total filler (K3) exhibited the most balanced performance, combining improved tensile strength with the lowest dielectric loss and a dielectric constant close to the targeted range. Higher filler loadings led to increased filler–filler interactions and microstructural heterogeneity, resulting in a more brittle mechanical response and elevated dielectric loss. Overall, the results demonstrate that the synergistic use of TESPI-modified glass and BaTiO_3_ fillers allows precise control over dielectric and mechanical properties in cyanate ester composites, highlighting their potential for high-frequency electronic and antenna-related applications.