Performance of epoxy hexagonal boron nitrate underfill materials: Single and mixed systems

Huai Man Ooi; Pei Leng Teh; Cheow Keat Yeoh; Chun Hong Voon; Nor Azura Abdul Rahim; Yun Ming Liew; Mohamad Syahmie Mohamad Rasidi; Mohamad Nur Fuadi Bin Pargi

doi:10.1515/epoly-2025-0007

Article Open Access

Performance of epoxy hexagonal boron nitrate underfill materials: Single and mixed systems

Huai Man Ooi , Pei Leng Teh , Cheow Keat Yeoh , Chun Hong Voon , Nor Azura Abdul Rahim , Yun Ming Liew , Mohamad Syahmie Mohamad Rasidi and Mohamad Nur Fuadi Bin Pargi

Published/Copyright: July 9, 2025

Published by

Become an author with De Gruyter Brill

Submit Manuscript Author Information

From the journal e-Polymers Volume 25 Issue 1

Abstract

The thermal mismatch between silicon chips, solder bumps, and substrates in electronic packaging raises reliability concerns, including solder joint failures and warping. To mitigate these issues, underfill materials with high thermal conductivity and a low coefficient of thermal expansion are essential. This study explores hexagonal boron nitride (h-BN) as a filler in epoxy composites, investigating both single and mixed filler systems. Epoxy resin and diethyltoluenediamine hardener were used, incorporating h-BN at 0–5 vol%. Optimal performance was achieved with 5 vol% of 70 nm h-BN, enhancing flexural strength (120 MPa) and thermal conductivity (0.17 W·m⁻¹·K). Mixed filler systems with nano- and micron-sized h-BN at varying ratios showed that a 25:75 ratio of 70 nm to 5 µm h-BN improved fracture toughness (7.21 MPa·m¹/²) and thermal stability. These results highlight the role of particle size distribution in optimizing epoxy/h-BN composites for electronic packaging.

Keywords: underfill; epoxy; hexagonal boron nitrate; single filler system; mixed filler system

1 Introduction

In recent years, the electronics industry has experienced significant growth due to rapid technological advancements and the increasing integration of electronics into everyday life. This surge has led to a greater demand for electronic packaging solutions, which are now crucial for ensuring the dependability, longevity, and functionality of devices like smart watches, laptops, and IoT systems. Technologies such as artificial intelligence, 5G, and the Internet of Things have further driven the need for innovative packaging that accommodates intricate circuitry while maintaining high performance standards (1).

Underfill technology has emerged as a key method in electronic packaging, especially for semiconductor devices with flip-chip designs. Underfill materials are applied to increase mechanical stability by filling the space between the chip and the substrate, thereby mitigating stresses caused by thermal cycling and mechanical shock (2). This method has become essential in fields such as automotive electronics, aerospace, and medical devices, where reliability is paramount for the durability and functionality of electronic equipment (3).

Epoxy resins, combined with fillers like silica, alumina, and microspheres, are commonly used in underfill applications due to their excellent adhesive properties, chemical resistance, and thermal stability (4). These fillers enhance the mechanical strength, thermal conductivity, and coefficient of thermal expansion (CTE) (5) matching of underfill materials, ensuring high performance in diverse operating conditions (6). Researchers are continually developing new formulations, including nanoparticle additives and alternative resins, to further improve the processability and sustainability of underfill solutions.

Hexagonal boron nitride (h-BN) has emerged as a promising filler due to its excellent thermal conductivity, which is essential for heat dissipation in electronic devices. In addition to its low dielectric constant and superior electrical insulation properties, BN’s hexagonal platelet structure provides reinforcement without sacrificing flow during encapsulation (7). BN also offers a lower CTE compared to other fillers, reducing the risk of failure from thermal cycling (8,9). Research indicates that the addition of h-BN can significantly improve thermal conductivity while maintaining electrical insulation. For instance, studies have demonstrated that composites with higher h-BN content exhibit enhanced thermal conductivity, with some achieving values up to 11.9 W·m⁻¹·K (10). Additionally, functionalization of h-BN fillers has been shown to further improve thermal conductivity and mechanical properties. Mehdipour et al. reported that electrospraying silane-treated h-BN onto carbon fiber surfaces and incorporating up to 20% h-BN into the epoxy matrix resulted in a 116% increase in thermal conductivity and enhanced mechanical properties (11). Similarly, functionalized h-BN modified through oxygenation and APTES treatment led to a 53% increase in thermal conductivity and a 57.14% improvement in toughness, though it introduced some brittleness (12).

Furthermore, incorporating micron-sized h-BN into epoxy composites facilitates the formation of a three-dimensional conductive network, achieving up to a 112% increase in thermal conductivity. This enhancement is complemented by improved mechanical properties, with flexural and tensile modulus increasing by up to 47%, making h-BN-reinforced epoxy composites suitable for advanced thermal management applications (13). Kirubakaran’s study optimized polymer hybrid composites with 5 wt% h-BN and 2 wt% banana fiber, achieving high thermal conductivity (1.03 W·m⁻¹·K) and electrical resistance (279.88 GΩ). The enhanced interfacial bonding and uniform filler dispersion make these composites ideal for heat shields, electrical insulation, and automotive applications (14).

Many studies have been conducted on systems with h-BN in epoxy resin, but limited research has focused on comparing single filler and mixed filler systems with different particle sizes. The aim of using a mixed filler system is to increase the filler packing density, thereby enhancing thermal conductivity and other properties (Tables 1–6).

Table 1

TGA results for epoxy/5 µm h-BN composites

Epoxy/5 µm h-BN	T _{on set} (°C)	T _max (°C)	T _{end set} (°C)	Weight loss (%)
1 vol%	348.9	373.5	409.4	77.91
5 vol%	349.3	373.5	408.1	70.01

Table 2

TGA results of epoxy/70 nm h-BN composites

Epoxy/70 nm h-BN	T _onset (°C)	T _max (°C)	T _{end set} (°C)	Weight loss (%)
1 vol%	347.8	379.8	407.5	74.62
5 vol%	349.2	372.0	409.9	73.30

Table 3

CTE values below and above T _g for epoxy/70 nm h-BN filled epoxy at 1 and 5 vol%

Epoxy/70 nm h-BN	CTE below T _g	CTE above T _g
1 vol%	48.8 × 10⁻⁶°C⁻¹	103 × 10⁻⁶°C⁻¹
5 vol%	44.1 × 10⁻⁶°C⁻¹	174 × 10⁻⁶°C⁻¹

Table 4

CTE values below and above T _g for epoxy/5 µm h-BN filled epoxy at 1 and 5 vol%

Epoxy/5 µm h-BN	CTE below T _g	CTE above T _g
1 vol%	58.0 × 10⁻⁶°C⁻¹	127.8 × 10⁻⁶°C⁻¹
5 vol%	50.8 × 10⁻⁶°C⁻¹	125.0 × 10⁻⁴°C⁻¹

Table 5

TGA results for the single and mixed filler system composites

Epoxy/70 nm/5 µm h-BN	T _onset (°C)	T _max (°C)	T _endset (°C)	Weight loss (%)
0:100	349.3	373.5	408.1	70.01
25:75	344.0	378.5	404.4	79.82
50:50	351.0	374.4	409.5	71.55
75:25	349.5	375.7	409.0	73.84
100:0	349.2	372.0	409.9	73.30

Table 6

CTE values below and above T _g for epoxy/70 nm/5 µm h-BN filled epoxy for the single and mixed filler system composites

Epoxy/70 nm/5 µm h-BN	CTE below T _g	CTE above T _g
0:100	40.8 × 10⁻⁶°C⁻¹	161 × 10⁻⁶°C⁻¹
25:75	41.3 × 10⁻⁶°C⁻¹	151 × 10⁻⁶°C⁻¹
50:50	42.6 × 10⁻⁶°C⁻¹	157 × 10⁻⁶°C⁻¹
75:25	42.5 × 10⁻⁶°C⁻¹	155 × 10⁻⁶°C⁻¹
100:0	44.1 × 10⁻⁶°C⁻¹	174 × 10⁻⁶°C⁻¹

2 Materials and methods

2.1 Material preparation

The composite formulation involved combining epoxy resin (Bisphenol A diglycidyl ether, DER331 grade, supplied by Euro Chemo-Pharma Sdn. Bhd, with 182–192 g epoxide equivalent weight) and diethyltoluenediamine hardener (was sourced from Shandong Aonuo New Material Company Ltd) with varying loadings of h-BN supplied by MK Impex Corp., Canada. For the single filler system, h-BN (5 µm) was added at 0, 1, 2, 3, 4, and 5 vol%. To ensure uniform dispersion and minimize agglomeration, the epoxy and h-BN were first mechanically stirred for 15 min, followed by 15 min of ultrasonic treatment. The ultrasonic process helped to break down particle clusters and enhance filler distribution within the matrix. After the ultrasonic treatment, the hardener was added and stirred for an additional 10 min to achieve a homogeneous mixture. The mixture was then degassed under vacuum for 10–15 min to remove air bubbles that could contribute to structural defects before being poured into molds (100 mm × 100 mm × 4 mm). Curing was performed at 120°C for 2 h to allow complete cross-linking. For the mixed filler system, a combination of 70 nm and 5 µm h-BN particles was used in ratios of 100:0, 75:25, 50:50, 25:75, and 0:100 (based on 5 vol% total filler content). The same mixing and curing procedures were followed to produce composite samples with different mixed filler compositions. This method enabled the preparation of composites with varying filler sizes and content for further analysis.

2.2 Flexural properties

Flexural testing measures the force required to bend a plastic beam, providing insight into the material’s stiffness and resistance to bending (15). During a three-point flexural test, the convex side of the specimen experiences tension, with the greatest stress on the outer fibers. In this study, flexural testing followed ASTM D790 standards using an Instron Universal Testing Machine 5569. The support span was set at 50 mm, with a crosshead speed of 2.38 mm·min⁻¹. Specimen dimensions were 60 mm × 12.7 mm × 3 mm. The flexural strength was calculated using Eq. 1.

(1) Flexural strength , δ f = 3 PL 2 bd 2

where δ _f = midpoint stress, MPa; P = Load at a specific point on the load-deflection curve; L = support span length, mm.

2.3 Fracture toughness

Fracture toughness refers to a brittle material’s ability to resist defect propagation under applied stress. It is commonly expressed as the critical stress intensity factor (K₁c), with higher K₁c values indicating greater fracture resistance and improved damage tolerance (16). The fracture toughness test was conducted using an Instron Universal Testing Machine 5569, following the ASTM D638 standard. Specimens measuring 60 mm × 12.7 mm × 3 mm were prepared with an initial 4 mm notch. The test was performed in tensile opening mode at a loading speed of 1 mm·min⁻¹. The calculation of fracture toughness is presented in Eq. 2.

(2)Fracture toughness, K c = F max π a Bw f a w

f a w = 1.99 − 0.41 a w + 18.7 a w 2 − 38.48 a w 3 + 53.85 a w 4

where K _c is the fracture toughness; F max is the maximum force in the force-deflection trace, N; a is the total notch length, m; B is the thickness of specimen, m; w is the width of specimen, m; and f a w is the geometric factor near unities that depend on details of specimen geometry.

2.4 Thermogravimetric analysis (TGA)

TGA is used to assess a material’s thermal stability and the proportion of volatile components by measuring weight changes as the sample is gradually heated (16,17). In this study, TGA was performed following the ASTM 1131 procedure using a Q500 TA Instrument. A 5 mg specimen was placed in a ceramic pan and heated from 50°C to 500°C at a rate of 10 °C·min⁻¹. throughout the process, the specimen was exposed to 50 mL of nitrogen per minute to ensure an inert atmosphere for accurate analysis.

2.5 Thermal conductivity

The thermal conductivity of the formulated underfill was tested using the RK-1 sensor with the KD2 Pro Thermal Properties Analyzer (8). Since the RK-1 sensor is suitable for solid samples, the epoxy/BN specimens were molded into cylindrical shapes with a length of 40 mm and a diameter of 30 mm, with a 4 mm hole drilled at the center. The instrument recorded thermal conductivity readings at fixed intervals of 15 s per reading to ensure accurate measurement.

2.6 Thermal expansion properties

The thermal expansion coefficient (CTE) of the composites was measured using a Linseis L75 Platinum Series horizontal dilatometer. The samples were heated from 30°C to 100°C at a rate of 5°C·min⁻¹ under a constant flow of argon gas. Disc-shaped samples were prepared with a thickness of 15–20 mm and a diameter of 10 mm. The CTE of each sample was calculated using the Eq. 3 to evaluate the thermal expansion across the temperature range.

(3) CTE , α = 1 L d L d T

where L is the measured length of the sample (mm); d L d T (µm °C⁻¹) is the change in the length of the sample per unit change in temperature.

2.7 Field emission scanning electron microscopy (FESEM)

FESEM uses a high-energy electron beam to produce signals from the specimen’s surface, revealing details about its shape, crystalline structure, orientation, and chemical composition (18). Before analysis, the specimen’s fracture surface was coated with palladium using a sputter coater. Microscopy was conducted with a NOVA NANOSEM 450 at activation voltages of 10 and 50 kV.

3 Results and discussion

3.1 Effect of different filler sizes on properties of epoxy composites

3.1.1 Flexural properties

Figure 1a shows that the flexural strength of 70 nm h-BN filled epoxy composites is higher than that of unfilled composites, with the highest strength observed at 5 vol% filler content. At this level, the h-BN particles are well dispersed in the epoxy matrix, allowing for effective stress transfer and improved bonding (19). Higher filler loadings can lead to poor dispersion due to increased Van der Waals forces, reducing flexural strength. Overall, the optimal filler content for 70 nm h-BN is 5 vol%, as it provides the best balance of dispersion and mechanical performance.

Figure 1

Flexural strength of epoxy composites (a) with 70 nm h-BN and (b) with 5 µm h-BN at various loadings.

The flexural strength of 5 µm h-BN filled epoxy composites decreases as filler content increases (Figure 1b). The reduction in flexural strength for 5 µm h-BN composites is attributed to limited filler-matrix interaction, as the larger particles receive insufficient epoxy resin coverage, weakening interfacial bonding. This inadequate coating leads to poor stress transfer, reducing the composite’s mechanical performance. Additionally, higher filler loading raises viscosity, making processing difficult and increasing void content, which further weakens the composites (20). Overall, 5 µm h-BN filled epoxy composites exhibit lower flexural strength compared to 70 nm h-BN composites, as the larger particle size leads to less surface area contact and more agglomerations, restricting the polymer’s molecular movement under load (21).

3.1.2 Fracture toughness

Figure 2a and b present the fracture toughness of epoxy composites filled with 70 nm and 5 µm h-BN, respectively. Both fillers exhibit increased toughness at 1–3 vol% loading, followed by a decline at higher contents. The 70 nm h-BN composites show slightly higher fracture toughness due to better dispersion and lower stress concentration, though their small size limits crack-arresting capability. In contrast, poor dispersion and weak interfacial bonding in 5 µm h-BN composites reduce toughness. These findings align with Kundie’s study, highlighting the critical influence of filler size on fracture toughness (21), with 1 vol% loading demonstrating optimal interfacial bonding.

Figure 2

Fracture toughness of (a) 70 nm h-BN filled epoxy composites and (b) 5 µm h-BN filled epoxy composites.

3.1.3 Thermal stability

Figure 3a shows the thermogram of epoxy/5 µm h-BN composites at 1 and 5 vol%. Both composites maintain thermal stability up to around 350°C. The initial decomposition temperature of the 5 vol% composite is slightly higher than the 1 vol% composite, indicating better thermal resistivity with increased filler content. Both composites experience major weight loss at 373.5°C, but the 5 vol% composite shows lower weight loss (70.01%) compared to the 1 vol% composite (77.91%).

Figure 3

Thermogram of (a) epoxy/5 µm h-BN composites and (b) epoxy/70 nm h-BN composites at 1 and 5 vol%.

Figure 3b shows the thermogram results for epoxy/70 nm h-BN composites. Both composites remain thermally stable up to about 350°C. The 5 vol% 70 nm h-BN composite has a slightly higher onset temperature than the 1 vol% composite, indicating improved thermal stability with higher filler content. The 5 vol% composite also shows 1.3% lower weight loss compared to the 1 vol% sample, confirming better thermal performance with increased filler content. By comparing different filler sizes of h-BN-incorporated epoxy composites, 5 vol% 5 µm filled epoxy has the lowest weight loss percentage and has the best thermal properties. In this case, the bigger filler size of h-BN introduced better thermal stability toward the epoxy matrix.

The trend shown in thermal stability is that the 70 nm h-BN high loading does not significantly enhance thermal stability, in contrast to that for epoxy/5 µm h-BN composites. This could be a result of the larger surface area of 70 nm h-BN particles, and while these could strengthen interfacial interactions with the epoxy matrix, they would also facilitate agglomeration at high filler loadings. This agglomeration can cause localized weaknesses or defects, which permit thermal decomposition at an earlier stage. Larger 5 µm h-BN particles, on the other hand, are unlikely to participate in agglomeration and able to form a more continuous heat-resistant skeleton, shielding the polymer from thermal degradation. The large particles might also aid in forming a more thermally stable char residue, further slowing down mass loss at elevated temperatures. These results indicate that while nanoparticle-scale h-BN provides good dispersibility at low loadings, high loadings could lead to instability as a result of particle aggregation.

3.1.4 Thermal conductivity

Figure 4 shows that the thermal conductivity of epoxy/h-BN composites increase with higher filler content. Smaller particles, like 70 nm h-BN, provide a greater surface area and more contact with the epoxy matrix, enhancing heat transport (8). As filler content increases, the viscosity rises, allowing particles to form better heat conduction pathways (20). Among all samples, 1 vol% of 5 µm h-BN has the lowest thermal conductivity, while 5 vol% of 70 nm h-BN has the highest. This demonstrates that smaller filler sizes lead to higher thermal conductivity due to the formation of more efficient conduction pathways.

Figure 4

Thermal conductivity of epoxy/h-BN at various filler contents.

3.1.5 Thermal expansion

Figure 5 compares the dimensional changes in 70 nm h-BN filled epoxy composites at 1 vol% and 5 vol% with increasing temperature, reflecting their CTE. The CTE increases for both filler contents as temperature rises. Above 100°C, the 5 vol% composite shows slightly less dimensional change than the 1 vol% composite, indicating that a higher filler content reduces thermal expansion. At lower temperatures, the dimensional changes are similar for both composites, but the difference becomes more pronounced as the temperature rises, due to stronger interaction between the epoxy matrix and h-BN particles. This reduction in CTE with higher filler content improves the composite’s dimensional stability, making h-BN a valuable filler for enhancing thermal properties.

Figure 5

CTE graph of 70 nm h-BN filled epoxy composites at different filler contents.

Figure 6 compares the dimensional changes in epoxy composites filled with 5 µm h-BN at 1 vol% and 5 vol% as temperature increases, reflecting their CTE. Both composites show increasing dimensional change with temperature, similar to nanoscale h-BN composites. The 1 vol% composite consistently exhibits a higher dimensional change across the temperature range compared to the 5 vol% composite, indicating that higher filler content and larger particle size improve the composite’s thermal dimensional stability and reduce its overall CTE. At lower temperatures (below 100°C), both composites show minimal differences in thermal expansion. However, above 100°C, the 5 vol% composite demonstrates significantly lower expansion, with the gap between the two curves becoming more pronounced. This suggests that the larger 5 µm h-BN particles enhance load transfer and provide better reinforcement to the epoxy matrix, effectively reducing thermal expansion. These findings highlight the importance of both filler content and particle size in optimizing the thermal properties of h-BN-filled epoxy composites, especially for applications requiring high-temperature stability.

Figure 6

CTE graph of 5 µm h-BN filled epoxy composites at different filler contents.

3.1.6 FESEM

Figure 7(a)–(c) presents micrographs of the 70 nm h-BN filled epoxy composite at 1,000×, 3,000×, and 10,000× magnifications, respectively. Figure 7(a) shows good distribution of the h-BN fillers in the epoxy matrix, though small gaps between the fillers and matrix are visible. Despite some agglomeration of the 70 nm h-BN fillers, the composite still exhibits high flexural strength at 5 vol% filler content, this is because the agglomerated h-BN particles are still embedded in the matrix (Figure 7(b) and (c)) due to its higher specific surface area with smaller particles size as compared to that of 5 µm h-BN epoxy system. Additionally, this filler content results in the highest thermal conductivity, likely due to the formation of effective conductive pathways for heat dissipation.

Figure 7

FESEM micrographs of epoxy/70 nm h-BN composite at 5 vol% (a) at 1,000× magnification, (b) at 3,000× magnification, and (c) at 10,000× magnification; epoxy/5 µm h-BN composite at 5 vol% (d) at 1,000× magnification, (e) at 3,000× magnification and (f) at 10,000× magnification, respectively.

Figure 7(d–f) presents micrographs of the 5 µm h-BN filled epoxy composite at 5 vol% at 1,000×, 3,000×, and 10,000× magnifications, respectively. Figure 7(d) shows the filler exhibiting moderate distribution due to its larger size. There is noticeable agglomeration of the filler particles, and gaps between the filler and the epoxy matrix indicate poor interfacial interaction. These gaps contribute to increased stress concentration, which explains the composite’s lower flexural strength at 5 vol% compared to the 2 and 3 vol% filler contents. The weaker filler-matrix interfacial adhesion causes the detachment of h-BN fillers and leaves the empty cavities on the surface of sample as seen in Figure 7(e) and (f), respectively.

In contrast, the 70 nm h-BN filled epoxy composite displays a better distribution within the epoxy matrix. The smaller particle size results in a higher specific contact area, allowing for more effective interaction with the matrix. This suggests improved interfacial contact between the filler and epoxy, leading to better homogeneity and enhanced mechanical properties.

3.2 Effect of mixed filler on properties of epoxy composites

3.2.1 Flexural properties

Figure 8 illustrates the relationship between the flexural strength of a composite material and various filler ratios of h-BN particles sized at 70 nm and 5 µm. Initially, at a 0:100 ratio of 70 nm to 5 µm h-BN, the composite shows a flexural strength of about 100 MPa, indicating a reasonable baseline strength with larger particles alone. However, as the ratio shifts to 25:75 with the addition of smaller 70 nm particles, the flexural strength drops to around 80 MPa. This decrease suggests that incorporating a small amount of smaller particles may not enhance the composite’s structural integrity, likely due to poor dispersion or insufficient bonding at this ratio.

Figure 8

Flexural strength of epoxy composites with mixed fillers.

Flexural strength significantly increases to about 90 MPa when the ratio reaches 50:50, indicating a synergistic effect where an equal mixture of both particle sizes improves material properties, possibly due to better stress transmission and particle packing within the matrix. When the ratio shifts to 75:25, the flexural strength returns to approximately 100 MPa, matching the original baseline value with larger particles. Notably, the maximum flexural strength of around 120 MPa occurs at a 100:0 ratio with only 70 nm h-BN particles. This substantial increase indicates that smaller particles provide greater reinforcement when used alone. The reduced particle size likely increases surface area, enhancing interfacial contact between the epoxy matrix and h-BN fillers, leading to improved mechanical properties and better load distribution.

In short, the results highlight that the size and ratio of h-BN fillers significantly influence the flexural strength of epoxy composites, with the highest benefit coming from smaller particles alone while a balanced mix of sizes also offers substantial advantages.

3.2.2 Fracture toughness

Figure 9 illustrates the fracture toughness of 70 nm/5 µm h-BN hybrid epoxy composites. The toughness decreases with an increasing 70 nm h-BN content, likely due to poor packing density and weak interfacial interaction, leading to more crack formation (22). Pure 5 µm h-BN epoxy shows the lowest toughness, as larger fillers distribute stress less effectively (19). The 25:75 ratio of 70 nm/5 µm h-BN is identified as optimal, achieving the highest toughness at 7.21 MPa·m¹/² by balancing particle sizes for improved dispersion and reduced crack-bridging.

Figure 9

Fracture toughness of mixed filler system epoxy composites.

3.2.3 Thermal stability

The thermal degradation behavior of epoxy composites loaded with varying ratios of 70 nm and 5 µm h-BN particles is illustrated in the TGA curves shown in Figure 10. All composites exhibit comparable thermal stability up to approximately 300°C, with minimal mass loss, indicating that the filler ratios do not significantly affect the epoxy matrix’s initial thermal resistance. However, thermal degradation begins to occur once the temperature exceeds 300°C. Notably, the composite with a 25:75 (70 nm:5 µm) ratio shows an earlier onset of mass loss compared to the other composites, suggesting slightly weaker thermal stability in this region. This reduced thermal stability may be attributed to suboptimal particle packing, where the specific ratio does not allow for an efficient thermal barrier formation (23). The weak interfacial interaction between fillers and the matrix at this composition could also create localized thermal stress points, promoting earlier thermal breakdown. Additionally, an imbalance in the synergistic effect between smaller and larger h-BN particles may contribute to the premature degradation, as effective heat dissipation relies on well-distributed and tightly packed fillers. The thermal breakdown of the epoxy matrix occurs across all composites between 300°C and 500°C, resulting in significant mass loss. Beyond 500°C, the mass reduction becomes slower, indicating further degradation and the formation of residual char.

Figure 10

Thermogram of epoxy/70 nm/5 µm h-BN composites at different mixed particle size ratios.

The thermal stability of epoxy composites is influenced by the filler-matrix interactions, particle dispersion, and the effectiveness of the filler in acting as a thermal barrier. In the case of epoxy/70 nm h-BN composites, increasing the filler loading does not significantly improve thermal stability. This could be due to the higher surface area of 70 nm particles, which may lead to stronger interfacial interactions with the epoxy matrix but also a higher tendency for agglomeration at high loadings. Such agglomeration can create localized defects or weak points that facilitate earlier thermal decomposition. Conversely, in epoxy/5 µm h-BN composites, larger h-BN particles are less prone to agglomeration and can form a more continuous heat-resistant network within the matrix. This helps to shield the polymer from thermal degradation more effectively, leading to an improved thermal stability trend as filler loading increases. Additionally, the larger particles may contribute to the formation of a more thermally stable char residue, further delaying mass loss at higher temperatures.

3.2.4 Thermal conductivity

Figure 11 illustrates the thermal conductivity of mixed systems of epoxy/h-BN composites. In epoxy composites, thermal conductivity is primarily governed by phonon transport, which can be disrupted by interfaces, particle dispersion, and matrix-filler interactions. Among all the samples, pure 70 nm h-BN exhibits the highest thermal conductivity. This suggests that smaller, well-dispersed particles create better conductive networks with minimal phonon scattering, leading to improved heat transfer (24). In contrast, the mixed particle size filler system with a 50:50 ratio of 70 nm to 5 µm h-BN filled epoxy composites displays the lowest thermal conductivity. This suggests that the homogeneous distribution in the 50:50 composite, with equal filler content of both particle sizes in the epoxy matrix, hinders the formation of effective heat conduction pathways. The 75:25 and 25:75 mixed 70 nm/5 µm h-BN epoxy composites show better thermal conductivity than the 50:50 composites. This is because having a higher proportion of one particle size leads to less homogeneous matrices, allowing the filler particles to clump together more, which facilitates the formation of additional conductive pathways.

Figure 11

Thermal conductivity of mixed filler system epoxy composites.

In summary, while the 50:50 ratio offers a balanced filler load, it results in lower thermal conductivity due to reduced particle interaction and clumping compared to ratios with a predominant particle size.

3.2.5 Thermal expansion

As shown in Figure 12, the CTE of all composites increase with temperature, which is a typical behavior for polymeric materials. However, the filler ratio significantly influences the extent of dimensional change. The composite with pure 70 nm particles at a 100:0 ratio (70 nm:5 µm) exhibits the greatest dimensional change across the temperature range, indicating a higher CTE. This suggests that larger 5 µm particles may be more effective at restricting the thermal expansion of the epoxy matrix compared to the smaller 70 nm particles.

Figure 12

CTE graph of different mixed filler ratios for 70 nm/5 µm h-BN filled epoxy composites.

In comparison, the mixed-ratio composites, specifically the 75:25 and 50:50 configurations, demonstrate less dimensional variation than the 100:0 composite. This confirmed that the inclusion of both larger and smaller h-BN particles appears to enhance the filler’s ability to limit the thermal expansion of the epoxy matrix. The 50:50 composite also shows a low CTE, slightly higher than the 75:25 composite. This suggests that while an equal proportion of both particle sizes provides good packing, the slightly increased presence of larger 5 µm particles in the 75:25 composite further enhances thermal constraint.

Additionally, the CTE of the 0:100 ratio composite, which consists solely of 5 µm particles, is lower than that of the 100:0 composite but not significantly different from the mixed filler composites. This indicates that while larger particles help reduce CTE, a combination of varying particle sizes may create a synergistic effect that enhances thermal dimensional stability even further.

The optimal filler ratio appears to be 75:25 (70 nm: 5 µm), as this composite shows less dimensional change than both the pure 70 nm (100:0) and pure 5 µm (0:100) particle composites, suggesting a lower CTE. The presence of more 70 nm particles likely improves particle dispersion and enhances matrix interaction, leading to better thermal dimensional stability. Overall, the findings indicate that blending different h-BN particle sizes within the epoxy matrix can yield a composite with improved thermal stability, characterized by lower CTE values. Particle size and distribution are critical factors, and mixed filler composites offer the best balance between dimensional stability and thermal conductivity (25).

3.2.6 FESEM

The FESEM micrographs in Figure 13a and b illustrates the filler distribution and microstructural features of the epoxy composite filled with a 75:25 ratio of 70 nm to 5 µm h-BN at various magnifications. At lower magnifications, the 5 µm particles are prominent and evenly distributed, while the smaller 70 nm particles are less visible, indicating proper mixing during composite production.

Figure 13

FESEM micrographs of 75% of 70 nm and 25% of 5 µm h-BN filled epoxy composite at (a) 1,000× magnification and (b) 10,000× magnification; 50% of 70 nm and 50% of 5 µm h-BN filled epoxy composite at (c) 1,000× magnification and (d) 10,000× magnification; 25% of 70 nm and 75% of 5 µm h-BN filled epoxy composite at (e) 1,000× magnification and (f) 10,000× magnification, respectively.

At higher magnifications, the 5 µm particles are surrounded by uniformly distributed 70 nm particles, showcasing strong adhesion between the fillers and the epoxy matrix with minimal gaps. This strong interfacial binding is crucial for effective load transfer and improved mechanical properties, as the high surface area of the smaller particles enhances interaction with the epoxy. However, at 10,000× magnification, some aggregations of the 70 nm particles are visible, which can create stress concentration areas and negatively affect the composite’s flexural modulus. The challenge remains to achieve uniform dispersion of nanoparticles due to their high surface energy, which often leads to clumping.

Figure 13c and d shows FESEM micrographs of an epoxy composite with a 50% ratio of 70 nm h-BN and 50% 5 µm h-BN at varying magnifications. At lower magnifications, the larger 5 µm particles indicate a uniform dispersion with minimal aggregation, while the smaller 70 nm particles surround them, suggesting effective mixing for consistent mechanical properties. At higher magnifications, the structure reveals well-dispersed 70 nm particles forming a network around the 5 µm particles, indicating strong interfacial adhesion. However, the 50:50 composite shows less continuous phase development than the 75:25 composite, which limits the formation of efficient heat conduction channels and results in the lowest thermal conductivity among the ratios.

Furthermore, the flexural strength of the 50:50 composite is lower than that of the 75:25 composite, as the higher concentration of 70 nm particles in the latter enhances interfacial bonding and load distribution. Thus, the lower thermal conductivity and flexural strength of the 50:50 composite stem from insufficient thermal channel formation and mechanical reinforcement.

Figure 13e and f presents FESEM micrographs of the epoxy composite with 25% 70 nm and 75% 5 µm h-BN particles at various magnifications. The images not only reveal generally uniform filler distribution but also show noticeable clustering of the larger 5 µm particles. Additionally, voids are apparent, likely caused by filler removal during the fracture toughness test, suggesting inadequate adhesion between the fillers and the epoxy matrix.

These voids may contribute to the observed decrease in flexural strength and modulus, indicating weak integration between the filler and matrix, which compromises structural integrity. Despite these mechanical drawbacks, the composite exhibits slightly better heat conductivity than the 50:50 ratio formulation. The presence of smaller particles may enhance thermal pathways, illustrating a trade-off between mechanical performance and thermal conductivity. This underscores the importance of optimizing filler dispersion and matrix adhesion for balanced performance.

4 Conclusion

In conclusion, this study on epoxy composites filled with h-BN fillers of varying sizes highlights key insights into their mechanical and thermal properties. It establishes a strong correlation between the size and volume percentage of h-BN fillers and the flexural strength of the composites. Specifically, 70 nm h-BN fillers at a 5 vol% ratio exhibit superior flexural strength (127.91 MPa) due to better dispersion and stress transfer. In contrast, larger 5 µm h-BN fillers result in reduced flexural strength (112.06 MPa) due to higher viscosity and poor interaction with the epoxy matrix. The optimal 5 vol% filler content enhances thermal conductivity and minimizes void content, ensuring effective stress distribution and improved rigidity.

For mixed filler composites, a 25:75 ratio of 70 nm to 5 µm h-BN particles is identified as the best combination, producing balanced improvements in mechanical and thermal properties. While this ratio shows slightly lower flexural strength (88.03 MPa) compared to the pure 5 µm h-BN composite, it exhibits the highest initial decomposition temperature (351.5°C), indicating enhanced thermal stability. Also, the 25:75 ratio 70 nm to 5 µm h-BN/epoxy composite shows the highest fracture toughness (7.21 MPa·m^1/2) compared to the other composites. The uniform distribution of fillers at this ratio underscores the importance of particle size distribution in optimizing the performance of epoxy composites for advanced material applications.

Acknowledgments

The authors would like to acknowledge Intel Electronics (Malaysia) Sdn. Bhd. for providing Intel Advanced Packaging Research Grant 2024 (9002-00135) to support this project.

Funding information: This material is based upon work supported by Intel Corporation under the funding award of the Advanced Packaging Research Grant 2024 cycle.
Author contributions: Huai Man Ooi: writing – original draft; Pei Leng Teh: review and editing the content and format; Cheow Keat Yeoh, Nor Azura Abdul Rahim, and Yun Ming Liew: review and checking of overall manuscript; Chun Hong Voon and Mohamad Syahmie Mohamad Rasidi: review and checking of content and grammer; Mohamad Nur Fuadi Bin Pargi: thermal analysis.
Conflict of interest: The authors declare no conflict of interest regarding this article.

References

(1) Wang H, Ma J, Yang Y, Gong M, Wang Q. A review of system-in-package technologies: application and reliability of advanced packaging. Micromachines. 2023;14(6):1149. 10.3390/mi14061149.Search in Google Scholar PubMed PubMed Central

(2) Nassiet V, Hassoune-Rhabbour B, Tramis O, Petit J-A. Electrical and electronics. In Adhesive bonding. Cambridge, United Kingdom: Elsevier; 2021. p. 719–61. 10.1016/B978-0-12- 819954-1.00025-3.Search in Google Scholar

(3) Fang K. Encapsulation process technology. In Encapsulation technologies for electronic applications. Vol. 1, Cambridge, United Kingdom: Elsevier; 2019. p. 123–81. 10.1016/B978-0-12-811978-5.00003-.Search in Google Scholar

(4) Sun Lee W, Yu J. Comparative study of thermally conductive fillers in underfill for the electronic components. Diamond Relat Mater. 2005;14(10):1647–53. 10.1016/j.diamond.2005.05.008.Search in Google Scholar

(5) Teh PL, Mariatti M, Wagiman ANR, Beh KS. Effect of curing agent on the properties of mineral silica filled epoxy composites. Polym Compos. 2007;29(1):27–36. 10.1002/pc.20345.Search in Google Scholar

(6) Dallaev R, Pisarenko T, Papezˇ N, Sadovsky´ P, Holcman V. A brief overview on epoxies in electronics: properties, applications, and modifications. Polymers. 2023;15(19):3964. 10.3390/polym15193964.Search in Google Scholar PubMed PubMed Central

(7) Mazumder MRH, Mathews LD, Mateti S, Salim NV, Parameswaranpillai J, Govindaraj P, et al. Boron nitride based polymer nanocomposites for heat dissipation and thermal management applications. Appl Mater Today. 2022;29:101672. 10.1016/j.apmt.2022.101672.Search in Google Scholar

(8) Abbasi S, Aravamudhan S. Effect of boron nitride (hBN) filler on thermal properties of underfill epoxy. 2017 16th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm); 2017, IEEE, p. 251–9. 10.1109/ITHERM.2017.7992479.Search in Google Scholar

(9) Danilov EA, Samoilov VM, Kaplan IM, Medvedeva EV, Stepashkin AA, Tcherdyntsev VV. Excellent thermal and dielectric properties of hexagonal boron nitride/phenolic resin bulk composite material for heatsink applications. J Compos Sci. 2023;7(7):291. 10.3390/jcs7070291.Search in Google Scholar

(10) Kirubakaran R, Salunke DR, Pitchumani SV, Gopalan V, Sampath A. Investigation of nano-HBN/natural fibers reinforced epoxy composites for thermal and electrical applications using GRA and ANFIS optimization methods. Polym Test. 2024;139:108561. 10.1016/j.polymertesting.2024.108561.Search in Google Scholar

(11) Mehdipour M, Doğan S, Al-Nadhari A, Sorayani Bafqi MS, Beylergil B, Yildiz M, et al. Influence of functionalized h-BN particle interphase and interface regulation with structural design on the directional thermal conductivity and mechanical performance of carbon fiber/epoxy composites. Compos Part A: Appl Sci Manuf. 2025;190:108708. 10.1016/j.compositesa.2025.108708.Search in Google Scholar

(12) Mehdipour M, Beylergil B, Yildiz M, Saner Okan B. Experimental validation and numerical modeling of interfacial effects in silanized hexagonal boron nitride (h-BN) reinforced epoxy composites by tailoring silane concentration. Compos Interfaces. 2025;1–21. 10.1080/09276440.2025.2460344.Search in Google Scholar

(13) Ozyigit S, Mehdipour M, Al-Nadhari A, Tabrizi AT, Dogan S, Dericiler K, et al. A comprehensive experimental study on the effects of hexagonal boron nitride particle size and loading ratio on thermal and mechanical performance in epoxy composites. J Compos Mater. 2024;58(13):1605–16. 10.1177/00219983241247910.Search in Google Scholar

(14) Kirubakaran R, Salunke DR, Pitchumani SV, Gopalan V, Sampath A. Development of HBN/natural fibres reinforced polymer composites using grey relation grade analysis for thermal and electrical applications. Funct Compos Struct. 2024a;6(2):025002. 10.1088/2631-6331/ad3edd.Search in Google Scholar

(15) Shrivastava A. Plastic properties and testing. In Introduction to plastics engineering. Cambridge, United Kingdom: Elsevier; 2018. p. 49–110. 10.1016/B978-0-323-39500-7.00003-4.Search in Google Scholar

(16) Mouritz AP. Fracture toughness properties of aerospace materials. In Introduction to aerospace materials. Cambridge, United Kingdom: Elsevier; 2012. p. 454–68. 10.1533/9780857095152.454.Search in Google Scholar

(17) Rajisha KR, Deepa B, Pothan LA, Thomas S. Thermomechanical and spectroscopic characterization of natural fibre composites. In Interface Engineering of Natural Fibre Composites for Maximum Performance. Cambridge, United Kingdom: Elsevier; 2011. p. 241–74. 10.1533/9780857092281.2.241.Search in Google Scholar

(18) Swapp, S. (2017). Scanning Electron Microscopy (SEM). Retrieved January 4, 2023, from https://serc.carleton.edu/research education/geochemsheets/techniques/SEM.html.Search in Google Scholar

(19) Navaneethakrishnan G, Karthikeyan T, Saravanan S, Selvam V. Influence of boron nitride on morphological, mechanical, thermal and wear characteristics of epoxy nanocomposites. Mater Res Innovations. 2020;24(5):257–62.10.1080/14328917.2019.1641346Search in Google Scholar

(20) Shafee MF, Jaafar M. Effect of boron nitride nano filler filled epoxy composites for underfill application. J Eng Sci. 2013;9:89–98.Search in Google Scholar

(21) Kundie F, Azhari CH, Muchtar A, Ahmad ZA. Effects of filler size on the mechanical properties of polymer-filled dental composites: a review of recent developments. J Phys Sci. 2018;29(1):141–65.10.21315/jps2018.29.1.10Search in Google Scholar

(22) Dungani R, Sumardi I, Alamsyah EM, Aditiawati P, Karliati T, Malik J, et al. A study on fracture toughness of nano-structured carbon black-filled epoxy composites. Polym Bull. 2021;78(12):6867–85. 10.1007/s00289-020-03444-5.Search in Google Scholar

(23) Lee Sanchez WA, Huang C-Y, Chen J-X, Soong Y-C, Chan Y-N, Chiou K-C, et al. Enhanced thermal conductivity of epoxy composites filled with Al2O3/boron nitride hybrids for underfill encapsulation materials. Polymers. 2021;13(1):147. 10.3390/polym13010147.Search in Google Scholar PubMed PubMed Central

(24) Lee E-S, Kang J-G, Kang M-K, Kim K-H, Park S-T, Kim YS, et al. High thermal conductivity of boron nitride filled epoxy composites prepared by tin solder nanoparticle decoration. Compos Part B: Eng. 2021;225:109264. 10.1016/j.compositesb.2021.109264.Search in Google Scholar

(25) Wang X, Niu X, Wang X, Qiu X, Wang L. Effects of filler distribution and interface thermal resistance on the thermal conductivity of composites filling with complex shaped fillers. Int J Therm Sci. 2021;160:106678. 10.1016/j.ijthermalsci.2020.106678.Search in Google Scholar

Received: 2025-02-10

Revised: 2025-04-05

Accepted: 2025-04-25

Published Online: 2025-07-09

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

https://doi.org/10.1515/epoly-2025-0007

Keywords for this article

underfill; epoxy; hexagonal boron nitrate; single filler system; mixed filler system

Creative Commons

BY 4.0