1. Verification of NanoDSC’s accuracy at high heating rates. Ge₂Sb₂Te₅, a widely researched phase-change material first commercialized by Intel, is chosen as a model system to evaluate whether NanoDSC is suitable for the fast-kinetic study of phase-change materials with reliable results.  NanoDSC quantifies the C_p(T), ΔH_c, and ΔH_m as a linear function of GST thickness (10 nm, 20 nm, and 40 nm). Their slopes align with the value measured in bulk samples, affirming NanoDSC as a reliable quantitative energy probe. In order to investigate the possible thermal lag in the system, a common source of instrumental error at fast scanning rates, we fabricated a sandwich-like GST-indium sample.  C_p(T) curves of indium melting are replicable and overlap perfectly with each other from 8.0×10³ to 1.0×10⁶ K/s. It demonstrates a high consistency of C_p(T) with negligible temperature shift in the indium melting (<2K) against different heating rates.  The result confirms the negligible thermal lag, repeatability of C_p(T), and stability of temperature calibration even at extremely high heating rates in NanoDSC.

2. Crystallication kinetics of GST. We further analyze the crystallization kinetic of GST using NanoDSC, which is the key material attribute for the fast data storage rate in PCM devices. NanoDSC reveals kinetic shifts of crystallization peaks with higher scanning rates. The maximum heating rate at ~10⁶ K/s is two orders of magnitude higher than any other calorimetry tool. It allows us to approach the crystallization kinetic proximate to the regime of fast switching in real devices (~10⁹ K/s). Assisted with KAS analysis, we concluded that the crystallization of GST is a single-step Arrhenius process with a constant activation energy of 2.36±0.14 eV. C_p(T) of GST crystallization with different thicknesses suggests that the process is dominated by growth with pre-matured nuclei at the interfaces. We numerically model the process and crystallization growth rate derived from NanoDSC are exactly within the range of that measured on PCM cells, and they all show similar Arrhenius behaviors. Assuming the growth of GST is diffusion-limited, based on the Stokes−Einstein relation, the viscosity of GST is derived which demonstrates a possible fragile-to-strong crossover at T~410 ^oC.

3. Melting of Phase-change superlattices. The study on GST paves our path for reliable NanoDSC study of other phase-change materials. Compared to Ge₂Sb₂Te₅, PCM devices made from 2D vdW superlattice (Ge₂Sb₂Te/ Sb₂Te₃) demonstrate 8 times lower Joule heating power during the melting cycle. To explore its mechanism, using NanoDSC, we uncovered a metastable transition of phase-change ST/GST SLs (65 nm thickness), which is not identified in the GST-ST phase diagram. C_p(T) of conditioned SL phase-change material demonstrates a sharp endothermic transition at ~380 ^oC. This transition is ~240 °C lower than the melting (~620 °C) of bulk GST, along with an 8× decrease in transition enthalpy. Such a transition provides original insights towards the low energy consumption and increased endurance of SL-based PCM devices. It should be noted that such a metastable transition requires strict operation temperature limits (<400 ^oC) in order to be reproducible as in real PCM devices. Otherwise, pulsing over 400 ^oC will lead to interlayer alloying and destruction of the superlattice structure. This is evident in C_p(T) evolution of the sequential high-T pulsing where metastable transition (T-380) is converted to bulk melting at ~600 ^oC. This is a seminal discovery in our work in the following two aspects: (1) we identify a metastable melting transition which is inherent in 2D vdW superlattices architecture. (2) we uncovered a material-level mechanism for failure in SL PCM devices and the optimal device operation guide (T<400 ^oC) is proposed accordingly to ensure maximum durability. Interestingly, such a transition at ~380 ^oC is absent in pristine GST (without vdW interfaces) but shows up in ST (with vdW interfaces). On the other hand, NanoDSC reveals that the low-T melting is always preceded by a shallow glass transition, indicating the existence of a small amount Te-rich glassy phase. It is thus proposed that the high density of vdW interfaces may facilitate the formation of Te-rich metastable phases in the superlattice and initiate the pre-melting of the lattice stacks.