Polyimide 기반 유·무기 복합 소재의 내열 및 방열 특성 = Thermal and Heat Dissipation Properties of Polyimide based Organic/Inorganic Composite Materials

다국어 초록 (Multilingual Abstract)

The rising integration density and output power of electronic devices have made heat accumulation a critical constraint on device reliability. Polyimide （PI） is an attractive matrix for thermal management, offering outstanding thermal stability and electrical insulation, yet its intrinsically low thermal conductivity restricts its use in heat-dissipation applications. The difficulty is that the necessary remedy works against itself: adding inorganic fillers raises thermal conductivity, but typically at the cost of heat resistance, insulation, or mechanical integrity.

Fifteen PI composite films were prepared with magnesium oxide （MgO） fixed at 10 wt% as a common base filler, combined with three secondary fillers chosen to differ in shape and electrical character: graphene oxide （GO, 1–4 wt%）, hexagonal boron nitride （BN, 2.23–20 wt%）, and zinc oxide （ZnO, 2.5–10 wt%）. All films were produced by simple physical stirring, without chemical surface modification or orientation control. This non-functionalized route was adopted deliberately: it reflects a low-cost, environmentally benign process that transfers directly to industry, and it isolates the contribution of the filler from confounding process variables.

FT-IR analysis confirmed the characteristic imide bands at 1,776, 1,708, 1,366, and 718 cm⁻¹ in every film, with no residual amic acid absorption near 1,660 cm⁻¹. All compositions were therefore equivalently imidized, and the matrix was chemically identical across the sample set—the premise on which the subsequent differences in properties can be attributed to the filler alone. OM, SEM, and EDS analyses distinguished the two-dimensional platelet morphology of BN and GO from the spherical morphology of ZnO. XRD showed that the integrated area of the BN （002） reflection increased with BN content—171, 451, 804, and 882 a.u. at 5, 10, 15, and 20 wt%—evidencing the progressive stacking and connection of BN platelets.

The thermal conductivity of the PI/MgO reference was 0.142 W m⁻¹ K⁻¹, and the enhancement followed the order BN 〉 GO 〉 ZnO. The BN composite rose monotonically to 0.218 W m⁻¹ K⁻¹ at 20 wt% （15.46 vol%）, a gain of 53.5%. The GO composite reached 0.160 W m⁻¹ K⁻¹ at 4 wt% （3.39 vol%） but could not be extended further, being limited by film formability; the ZnO composite peaked at 0.154 W m⁻¹ K⁻¹ at 7.5 wt% and then fell to 0.131 W m⁻¹ K⁻¹ at 10 wt%. Compared at an equal volume fraction of approximately 2.6 vol%, the order BN （0.152） 〉 GO （0.143） 〉 ZnO （0.131） persisted. Shape, not loading, therefore governs pathway formation: platelet fillers meeting face to face construct conduction routes more efficiently than spherical fillers confined to point contacts. Low density and superior film formability allowed BN alone to extend these routes to high volume fractions.

Tg-DTA analysis revealed that the BN composite exhibits a threshold behaviour with respect to content. At 2.23 and 3.70 wt%, T₅% fell to 526.1 and 530.9 ℃, below the reference of 553.9 ℃: sparse platelets cannot form a continuous barrier network, and the filler–matrix interface instead acts as a local defect. Above 5 wt% the heat resistance recovered and then improved steadily—548.3 ℃ at 5 wt%, 551.1 ℃ at 10 wt%, 553.7 ℃ at 15 wt%, and 561.2 ℃ at 20 wt%, the highest of all fifteen films. The GO composite remained comparable to the reference throughout （547.0–557.2 ℃; char yield 81.17–82.71%）, whereas the spherical ZnO composite was markedly inferior （478.1–503.9 ℃; char yield 73.35–79.58%）.

Dynamic mechanical analysis showed that the platelet fillers substantially reinforced the glassy storage modulus. The BN composite peaked at 4,414 MPa at 15 wt%—roughly three times the reference of 1,491 MPa—and reached 3,819 MPa even at 2.23 wt%. The GO composite attained its maximum of 3,077 MPa at 3 wt%, where it simultaneously recorded the highest glass transition temperature （427.3 ℃） and the lowest tan δ peak （0.154） of the entire set; at this loading GO forms its greatest effective interfacial area and most strongly restricts segmental chain motion. The spherical ZnO composite, by contrast, remained between 1,556 and 2,101 MPa, and its glass transition temperature declined monotonically from 405.2 ℃ at 2.5 wt% to 388.6 ℃ at 10 wt%.

Break load was adopted as the primary tensile index. Laboratory-scale hand casting produced film thicknesses of 0.02 to 0.07 mm, so cross-section–normalized strength would reflect casting deviation rather than intrinsic material behaviour. Against the reference of 33.40 N, the BN composite rose monotonically above 5 wt% （R² = 0.82） to 45.08 ± 2.39 N at 20 wt%—135.0% of the reference—with the highest elongation at break of all fifteen films （7.10%） and the smallest scatter （CV 5.3%）. The ZnO composite fell steadily by 1.51 N per wt% between 2.5 and 7.5 wt% （R² = 0.96） and remained at 55–86% of the reference throughout. Since three to four specimens were measured per composition and the scatter reached 35–44% of the mean in some cases, these results are offered as supporting evidence rather than as a quantitative ranking.

Heat dissipation, heat resistance, and mechanical reliability thus followed one and the same order—platelet fillers （BN, GO） 〉 spherical filler （ZnO）—and one factor was common to all three: the shape of the filler and the contact geometry it produces. The continuous, face-to-face pathways formed by platelet fillers serve simultaneously as conduction routes, as barriers retarding the diffusion of decomposition products, and as media for stress transfer; a single mechanism governs all three properties. Among the fifteen films, BN 20 wt% met these demands at the highest level, combining a thermal conductivity of 0.218 W m⁻¹ K⁻¹ （+53.5%）, a T₅% of 561.2 ℃, a break load of 45.08 N （135.0%）, and an elongation at break of 7.10%—the highest values in the set for each of the latter three. Together with the intrinsic electrical insulation of BN, which the electrically conductive GO cannot provide, these results identify BN as the secondary filler that best satisfies the simultaneous demands of heat dissipation, heat resistance, and insulation.

The contribution of this work is to establish the correlation among filler shape, volume fraction, and properties as a single design principle, achieved under simple physical stirring alone. Where much previous work on heat-dissipation composites has depended on surface modification or high-degree orientation—routes that limit industrial scalability—this study quantifies both the attainable performance and its governing factor under a low-cost, environmentally benign process, and so offers practical guidelines for the design of industrially applicable heat-dissipation composites.

Keywords: polyimide, organic/inorganic composite, heat dissipation, thermal conductivity, boron nitride, filler shape, volume fraction, heat resistance

목차 (Table of Contents)

• 1. 서 론 20
• 2. 이론적 배경 27
• 2.1 폴리이미드 27
• 2.2 이미드화 메커니즘 30
• 2.3 고분자 복합소재의 열전달 메커니즘 32
• 2.4 필러의 종류와 특성 39
• 2.5 복합소재의 내열성·절연성 및 기계적 특성 46
• 2.6 필러 구조와 공정적 한계 49
• 3. 실험 및 방법 52
• 3.1 시료 및 시약 52
• 3.2 시료 설계 56
• 3.3 복합필름의 제조 60
• 3.4 특성 분석 66
• 4. 결과 및 고찰 75
• 4.1 화학적 구조 및 이미드화 분석（FT-IR） 75
• 4.2 표면 분산 형태 분석（Optical Microscopy, OM） 84
• 4.3 미세조직 분석（Scanning Electron Microscopy, SEM） 92
• 4.4 원소 조성 분석（Energy-Dispersive X-ray Spectroscopy, EDS） 104
• 4.5 결정구조 분석（X-ray Diffraction, XRD） 140
• 4.6 내열 특성-열안정성 분석（TG-DTA） 149
• 4.7 동적기계분석（DMA） 155
• 4.8 방열 특성—열전도도（ASTM D5470） 160
• 4.9 인장 특성 분석（UTM） 170
• 5. 결 론 176
• 참고문헌 183
• Abstract 204