Objective The formation of cellulose nanocrystals (CNCs) is conventionally attributed to the preferential degradation and removal of amorphous regions during acid hydrolysis. However, this prevailing view fails to account for the spontaneous recrystallization of amorphous cellulose observed under aqueous or hydrothermal conditions. To address this gap, this study focuses on the crystal facet reconstruction of amorphous cellulose across diverse media. We aim to elucidate the evolving proportion of specific crystallographic planes during recrystallization and to clarify how medium conditions govern polymorphic transformations and the restoration sequence of crystal facets.

Method First, the evolution of specific crystal facets under varying ball-milling durations was compared to analyze the selective destructuring of the cellulose crystalline architecture. Subsequently, using amorphous cellulose as the starting material, its recrystallization behavior was investigated in both hydrothermal and oxalic acid systems. The dynamic evolution of the (200), (1 \bar 1 0 ), (110), and (004) crystallographic planes was systematically tracked via Gaussian multi-peak fitting of X-ray diffraction (XRD) patterns combined with Fourier transform infrared (FTIR) spectroscopy.

Result Ball milling significantly broadened and attenuated the characteristic peaks of Cellulose I, reducing the relative crystallinity from 80.1% to 28.9%. Crystallographic analysis revealed that the (1 \bar 1 0 ) plane degraded more rapidly than the (110) plane, while the pronounced broadening of the (004) plane indicated a facet-selective deconstruction dominated by the preferential disruption of lateral hydrogen-bonding networks. Subsequently, the milled samples were subjected to separate treatments at 100 °C for 2 h—either under hydrothermal conditions or in an aqueous oxalic acid solution—which restored the relative crystallinity to 65.5% and 68.2%, respectively, and induced the emergence of characteristic peaks of the Cellulose II. Quantitative analysis of facet proportions demonstrated that recrystallization did not restore the original crystalline architecture; rather, it established a new equilibrium between Cellulose I and II. Specifically, the hydrothermal system preferentially restored the (200) plane, whereas the oxalic acid system facilitated the recovery of the (004) plane at lower temperatures. At elevated temperatures, however, the facet distributions in both systems converged toward a typical Cellulose II profile.

Conclusion In conclusion, ball-milling-induced amorphization altered the proportional composition of crystallographic planes, revealing that recrystallization essentially follows a “selective disruption–proportional redistribution–new equilibrium formation” pathway of crystal facet reconstruction. While the reaction medium governed the restoration sequence of specific facets by modulating chain segment swelling and hydrogen-bond reformation, it did not alter the ultimate direction of the polymorphic transition. These findings suggest that, in addition to the conventional “amorphous region removal” pathway, CNC formation can also originate from the medium-induced rearrangement of disordered chains. This work provides a theoretical basis for the low-acid, environmentally benign preparation of CNCs.