Probing effect of strain field on Ca diffusion in BaZrO₃-doped YBa₂Cu₃O_7-x /Ca_0.3Y_0.7Ba₂Cu₃O_7-x multilayer nanocomposite films

Recently, we reported that incorporating thin Ca_0.3Y_0.7Ba₂Cu₃O_7-x spacer layers in BaZrO₃/ YBa₂Cu₃O_7-x (BZO/YBCO) films in multilayers (ML) could lead to a coherent BZO/YBCO interface due to reduced lattice mismatch of ∼1.4% and hence improved pinning by the BZO 1D artificial pinning centers (BZO 1D-APCs). It was hypothesized that the highly strained YBCO columns around the BZO 1D-APCs due to the larger lattice constant of BZO than YBCO's in c-axis by 7.7% provide channels to facilitate Ca diffusion more preferably near the BZO/YBCO interface, leading to an enlarged YBCO's c-axis lattice constant as a consequence of the Ca/Cu substitution on the YBCO's Cu-O planes. Confirmation of this hypothesis is hence important to understand the mechanism of Ca diffusion in BZO/YBCO ML films for further enhancement of pinning efficiency of the BZO 1D-APCs. Motivated by this, this work presents a comparative study of two sets of BZO/YBCO ML films: singly-doped (SD-ML, BZO only) and the doubly-doped (DD-ML, BZO + Y₂O₃). The modulated strain field in the SD-ML films is in contrast to the non-modulated strain in the DD-ML samples as confirmed by transmission electron microscopy. Interestingly, improved pinning by >four times of the reference SD single-layer (SD-SL) film’ was observed in the SD-ML samples even when the constituent BZO/YBCO layer thickness was varied between 50-100 nm, indicating that the Ca diffusion along the strained BZO/YBCO interface and the Ca/Cu substitution on YBCO's Cu-O planes indeed correlate with the modulated strain field. In contrast, pinning degrades by >twice in DD-ML samples with respect to their DD-SL counterpart's. This result therefore suggests that strained BZO/YBCO interfaces serve as channels for Ca diffusion in SD-ML, which leads to the coherent BZO/YBCO interfaces and hence improved pinning of BZO 1D-APCs. This effect diminishes when such a modulated strain field is removed in the DD-ML samples.