Experimental and theoretical study on the thermal conductivity of PA6/POE/AlN nanocomposites prepared by in-situ and melt polymerization methods

Polyamide 6/polyolefin elastomer (PA6/POE) nanocomposites containing \(\:5-20\:wt{%}\) aluminium nitride (AlN) were prepared by two routes, in-situ anionic ring-opening polymerisation of ε-caprolactam and conventional melt compounding, in order to separate the effect of the processing route on filler dispersion, phase morphology and heat transport. Field-emission scanning electron microscopy showed that the in-situ route extends the POE phase into stable fibrils, whereas melt compounding leaves it as discrete droplets; image analysis of the filler-free references gives a median equivalent domain diameter of 2.8 μm for the in-situ blend against 3.5 μm for the melt-compounded blend. Oscillatory rheometry confirmed the difference: only the samples in which AlN is localised at the PA6/POE interface develop a low-frequency plateau in the storage modulus and a solid-like response. Thermal conductivity increased monotonically up to 15 \(\:wt{%}\) AlN, reaching a maximum of 3.06 \(\:W · {m}⁻¹ · {K}⁻¹\) for the in-situ material, seven times the value of the unfilled blend and 7% above the melt-compounded equivalent, and then fell to 1.36 W·m − 1 ·K − 1 at 20 \(\:wt{%}\) as the thermal diffusivity collapsed. Model analysis was carried out over the complete composition range. Between 0 and 15 \(\:wt{%}\) the Agari model describes the data well (R 2 = 0.95 in the logarithmic space in which it was fitted, 0.89 on a linear scale), whereas the Nielsen and Bruggeman models underestimate the measured conductivity roughly fivefold and return negative coefficients of determination. Across the full 0-20 \(\:wt{%}\) range the Agari value falls to R 2 = 0.51, because every model considered increases monotonically with filler content and none can reproduce the reversal above 15 \(\:wt{%}\) . That structural limitation, rather than scatter in the data, is what the full-range statistics record. Interfacial filler localisation and fibrillar networking are therefore confirmed as a workable route to thermally conductive polyamide compounds, with 15 \(\:wt{%}\) AlN identified as the practical optimum rather than the highest loading examined.