Design of shape memory alloys with enhanced thermal management properties via adaptively constrained multi-objective optimization

Shape memory alloys (SMAs) as solid–solid phase change materials are promising candidates for addressing thermal loads in electronic devices. High thermal energy storage (TES) performance and narrow overall transformation range (OTR) are desirable for rapidly mitigating excess heat in devices with confined spaces and narrow operating temperature windows. However, achieving a combination of high TES performance and a narrow OTR remains challenging due to the inherently conflicting nature of properties and the vast compositional design space. To address this, we propose an adaptively constrained multi-objective optimization strategy that evaluate the expected hypervolume gain of candidates within a constrained preferred region. Through visualization on a mathematical function case and comparative experimental exploration on SMAs, this constrained strategy demonstrates significantly enhanced efficiency in guiding the search toward favorable property trade-offs. With the constrained strategy, all 10 synthesized alloys across three experimental iterations exhibit high TES performance and narrow OTR. Especially, alloy Ti_50.2Ni_43.6Cu_5.8Al_0.4 displays a high TES performance of 2690 × 10⁶ J² K⁻¹ s⁻¹ m⁻⁴, which is over 6 times higher than commercial organic PCMs, and 2∼3 times larger than the best-performing TiNiCu (B2–B19) alloys reported in the literature for electronic thermal management. Its narrow OTR (26.81 °C) lies well within the acceptable operating temperature range of electronics. Moreover, it demonstrates good thermal stability after rolling, with a slight shift of 0.18 K in transformation temperature and a decay of only 0.15 J/g in latent heat after 40 cycles.