Atomic-scale imaging reveals how vacancy chemistry and lattice strain shape the birth and behavior of copper nanoparticles inside perovskite oxides.
Paper: Exsolution of metal nanoparticles by control of point defects. AI-generated abstract conceptual image created using ChatGPT/OpenAI
*Important notice: This news reports on an unedited version of an accepted paper and is awaiting final editing. Therefore, the paper should not be regarded as conclusive or treated as established information.
In a recent article in Nature Communications, researchers describe the nanoscale mechanisms of metal nanoparticle exsolution from perovskite oxides by controlling point defects, clarifying how to tune the process for catalytic applications.
Defect Chemistry and Exsolution
Exsolution is a process in which metal nanoparticles precipitate from metal oxides through redox reactions, yielding materials with promising catalytic and electrochemical properties.
This phenomenon produces nanoparticles socketed onto oxide substrates, improving stability against agglomeration and creating strong catalytic interfaces at triple-phase boundaries. Exsolution is important for developing nanoscale heterostructures and regenerable catalysts, but a fully coherent understanding of the underlying mechanisms has been lacking.
Exsolution differs from phase decomposition because it retains the host oxide’s crystal structure, though it can still cause secondary-phase formation or surface cation enrichment in some systems.
Nanoparticle generation and behavior are linked to point-defect chemistry in the oxide matrix, which can be deliberately engineered.
Tuning A-site deficiencies in perovskite structures is a known strategy to facilitate exsolution by enabling vacancy consumption that leads to void formation, allowing nanoparticles to become socketed either at the surface or within the bulk as embedded endoparticles.
The authors describe the process with three coupled point-defect reactions: reduction of the host oxide, reduction and exsolution of copper, and annihilation of host unit cells through the reverse Schottky reaction. Their model predicts that the parent composition has enough oxygen vacancies to support about 50% copper exsolution via this pathway, while trivalent B-site acceptors can raise the theoretical limit.
Acceptor Substitution and Experimental Analysis
This study uses acceptor-substituted perovskites based on La0.2Ca0.7Ti0.95Cu0.05O3−δ, where Ga, Al, and Sc partially substitute for Ti on the B-site to tune point-defect concentrations and lattice parameters. The materials underwent controlled reductive treatments at 750–800 °C. In situ XRD and XANES used 4% H2 at 800 °C for 2 h, while ex situ XANES and nanoscale imaging examined samples reduced in 5% H2 at 750 °C for 24 h.
Synchrotron X-ray diffraction (XRD) and X-ray absorption near-edge structure spectroscopy (XANES) quantified exsolved copper metal from crystallographic signatures and oxidation states.
Scanning transmission electron microscopy (STEM) with annular-dark field (ADF) and annular-bright field (ABF) imaging, energy-dispersive X-ray spectroscopy (EDS), and electron energy loss spectroscopy (EELS) enabled atomic-scale visualization of the nanoparticles, their sizes, distributions, crystallographic relationships, and local chemical composition.
Thermogravimetric analysis monitored mass and oxygen stoichiometry changes during exsolution. Atomic column mapping, combined with line-profile refinement, clarified the relationship between copper nanoparticles and oxygen columns in the host lattice, helping the researchers assess coherency and strain.
Together, these methods let the researchers examine how point-defect engineering affects nucleation density, particle size, anchoring mechanisms, and lattice strain at the nanoscale.
Exsolution Yield and Nanoparticle Behavior
Gallium and aluminum acceptor substitution increased copper exsolution compared with the parent compound, as shown by stronger metallic Cu(111) reflections in XRD and Cu(0) features in XANES spectra. Both also produced faster initial exsolution. XRD gave lower absolute estimates than XANES because diffraction peaks from nanosized particles were broadened.
STEM imaging revealed that nanoparticles in acceptor-doped samples were smaller, around 2 nm for Ga substitution versus about 5.5 nm in the parent material. The smaller particle sizes are consistent with higher nucleation density and strong anchoring, which limit particle diffusion and coalescence.
Under the 750 °C, 24-hour imaging conditions, the nanoparticles were found predominantly as endoparticles embedded within the host lattice.
Atomic-resolution ADF and ABF images showed copper columns aligned with oxygen columns near the particle center. Toward the particle edges, the columns became progressively misaligned because the copper and perovskite lattices have different dimensions. This supports a coherent crystallographic relationship with the host and a face-centered cubic structure consistent with metallic copper.
These observations support a solid-state diffusion and nucleation mechanism directly from point defects rather than phase decomposition. The paper also describes later diffusion and coalescence of endoparticles, including their emergence at surfaces.
Tensile strain from the lattice mismatch between the smaller copper lattice and the larger pseudocubic perovskite lattice affected anchoring strength and particle coalescence.
After two hours at 800 °C, both aluminum- and gallium-substituted samples contained large surface nanoparticles of about 100 nm.
The Ga sample also retained nanometric particles within the bulk, yielding both small embedded and larger surface particles, whereas the Al sample showed no endoparticles under these conditions.
Strain also created a thermodynamic and kinetic barrier to further exsolution at larger lattice mismatch. Scandium substitution enlarged the pseudocubic lattice parameter and increased interfacial strain.
During the first two hours at 800 °C, the Sc sample exsolved copper faster than the parent sample. After 24 hours at 750 °C, XANES measured a lower exsolved copper fraction.
Ex situ XRD showed the opposite trend, with higher exsolution yield as scandium content increased, which the authors associate with larger surface nanoparticles. The apparently contrasting results reflect differences in treatment conditions and in the particle populations detected by the two methods.
Measurements showed that the oxygen occupancy of the remaining host oxide increased upon exsolution, while the sample lost oxygen during reduction.
The sample continued to release oxygen rather than absorb it. Two oxygen vacancies are consumed for each copper atom exsolved in the proposed defect reaction. At the same time, host unit cells are removed, so the remaining perovskite contains a higher proportion of occupied oxygen sites.
Atomic-scale EELS measurements provided structural evidence for this proposed unit-cell annihilation. The elemental maps showed localized decreases in Ca, Ti, and O signals at an embedded copper nanoparticle while the surrounding host structure remained intact.
Perovskite unit-cell annihilation creates voids where endoparticles can nucleate and remain embedded, linking the defect reaction to the socketed geometry of exsolved nanoparticles.
Thermodynamics and Strain Effects
The study shows that controlled manipulation of point defects through acceptor substitution can increase copper exsolution and nucleation density, while lattice mismatch influences particle size, shape, anchoring strength, and the extent to which strain limits further exsolution.
The work provides a defect-chemical basis for designing exsolution systems with controlled particle populations. It does not directly test catalytic activity or long-term durability, so claims about catalyst performance remain prospective. The findings may help guide the design of catalysts and nanoscale heterostructures in which particle size, distribution, and anchoring can be controlled more precisely.