Tailoring atomic diffusion for in situ fabrication of different heterostructures

Hetero-nanostructures combine multiple functional components to achieve synergistic properties, with one-dimensional (1D) forms enabling applications in energy storage, catalysis, sensing, and optoelectronics via heterointerface effects. Morphologies typically include core-shell and segmented structures, often dictated by the synthesis route. Diffusion-based methods are attractive for their simplicity and potential for interface/morphology control, but achieving tailored atomic diffusion in individual nanostructures remains challenging. This study investigates whether electromigration-driven, in situ solid-state diffusion under bias in a TEM can reliably tailor diffusion pathways to selectively fabricate either core-shell or segmented Ag–Te heterostructures, thereby addressing bottlenecks in controlled nanoscale diffusion mechanisms for integrated nanodevices.

The paper surveys three main categories for synthesizing 1D heterostructures: heterogeneous nucleation and growth, diffusion and deposition, and melt-casting/electrospinning. Diffusion methods (e.g., heterogeneous diffusion, Kirkendall effect, ion exchange, solid-state phase transformation) offer improved interface and morphology control. Prior works include solution-phase in situ diffusion growth of Fe(OH)₃@α-MoO₃ core-shell nanorods, solid-state diffusion for Ta-doped TiO₂, and electrically driven in situ cation exchange to form CdS–Cu₂S core-shell nanowires. These advances underscore the promise of diffusion approaches but also highlight remaining gaps in controlling atomic/ionic diffusion and mechanisms at the nanoscale for individual heterostructures.

Silver (Ag) and tellurium (Te) nanowires (NWs) were synthesized via solvothermal methods. Ag NWs: ethylene glycol with PVP was added to 0.1 M AgNO₃, transferred to a Teflon-lined autoclave, heated at 160 °C for 2.5 h, then washed. Te NWs: TeO₂, PVP, NH₃·H₂O, and hydrazine monohydrate were dissolved in water, sealed in an autoclave at 200 °C for 16 h, then washed with ethanol and water. For in situ TEM experiments, Te NWs were deposited on a half Cu grid mounted on a gold wire; W tips were etched and plasma-cleaned, then dipped into Ag NW suspension so Ag NWs adhered to the tip. Inside an aberration-corrected TEM (FEI Titan 80–300 kV) with a TEM-STM holder, an Ag NW was brought into contact with a single Te NW. Bias voltages of controlled polarity and magnitude were applied to drive atomic diffusion while imaging (TEM/STEM), recording videos, I–V measurements, and collecting HRTEM, FFT, HAADF-STEM, and EDX elemental mappings. Growth kinetics were monitored from time-resolved image series. A heat-conduction model for a 1D conductor under bias (considering Joule heating, contact resistance, and differing thermal conductivities of Ag and Te) was used to estimate temperature distributions and their effects on diffusion pathways. Characterization used TEM/STEM, HRTEM with FFTs to identify phases and orientation relationships, and EDX for radial elemental profiles to distinguish surface versus bulk diffusion.

• Polarity-controlled morphology: Applying a positive voltage to the Ag NW (relative to Te) led to surface diffusion of Te into Ag and formation of Ag₂Te–Ag core-shell NWs; reversing the polarity resulted in bulk diffusion of Ag into Te and formation of segmented Ag₂Te–Te heterostructures.
• Core–shell formation under +0.3 V: Te atoms migrated along the Ag NW surface, nucleating Ag₂Te hillocks that coalesced into a continuous shell. The reaction 2Ag + Te → Ag₂Te is spontaneous (ΔG = −45.4 kJ mol⁻¹). Growth proceeded via two stages (nucleation/growth of hillocks, then coalescence/thickening). With stable contact and constant bias, shell thickness increased linearly with time at ~0.038 nm s⁻¹. Orientation relationship identified as Ag₂Te{123}//Ag(111). EDX radial profiles showed uniform Te in the shell and Gaussian-distributed Ag in the core, consistent with surface diffusion of Te.
• Segmented heterostructure under negative bias (e.g., −0.8 V): Ag atoms electromigrated into Te NW, initially without obvious structural change; after ~80 s a clear Ag₂Te phase formed with diameter expansion by ~15% (~30% volume increase), producing an Ag₂Te–Te segmented NW with a transition zone. The phase transformation front advanced at ~2 nm s⁻¹ under constant current, while the Ag diffusion front slowed over time (~8 nm s⁻¹ to ~3 nm s⁻¹) due to increasing diffusion length. EDX in the transition zone showed Gaussian radial distributions for both Ag and Te, indicating bulk diffusion. A very weak Te signal was detectable near the Ag side (Te:Ag ≈ 1:(73.31 ± 35.27)), suggesting minor thermal diffusion.
• Electrical characteristics: Formation of a semiconducting Ag₂Te layer led to a Schottky-type contact at the Ag₂Te–Ag interface, changing the I–V behavior from ohmic (linear) to S-shaped.
• Mechanism: Electromigration (electron-wind force) drives atoms opposite to the electric field direction; not an electrochemical process (bias < standard electrode potentials of Ag and Te). Electron-beam effects were negligible and isotropy expected from beam-induced diffusion was not observed.
• Temperature effects (Joule heating): Modeling and parameter estimates suggest large temperature rises on the Te side (>300 K, approaching Te melting temperature) due to low thermal conductivity and significant contact resistance, enabling bulk diffusion of Ag in Te. On the Ag side, temperature rise is minimal (<10⁻²–10 K), so surface diffusion of Te dominates, determining the final morphologies.
• Generality and tunability: Reaction speeds depend on current density and NW dimensions (thinner NWs react faster). Growth and transformation rates are tunable via applied bias and contact conditions.

The findings demonstrate that the direction and pathway of atomic diffusion in Ag–Te nanowire couples can be deterministically controlled by electrical bias polarity and local temperature, directly addressing the challenge of tailoring solid-state diffusion at the single-nanostructure level. Electromigration, rather than electrochemistry or electron-beam effects, governs atom transport: atoms migrate opposite to the electric field due to electron-wind forces, and phase transformation to Ag₂Te proceeds spontaneously once the local Ag:Te ratio is sufficient. Joule heating, amplified by high contact resistance and the low thermal conductivity of Te, sets the diffusion mode: elevated Te-side temperatures permit bulk diffusion of Ag in Te, forming segmented Ag₂Te–Te structures; low temperature rise on the Ag side favors surface diffusion of Te on Ag, leading to core–shell Ag₂Te–Ag structures. The polarity-dependent diffusion paths correlate with measured growth kinetics, I–V changes, and EDX radial profiles. These insights clarify the micro-mechanisms of solid–solid diffusion reactions under bias and suggest practical routes to tailor heterostructure morphologies by adjusting current, contact quality, and heat dissipation.

An in situ, electrically driven solid–solid diffusion strategy enables selective fabrication of individual Ag–Te heteronanostructures with distinct morphologies: Te surface diffusion on Ag produces Ag₂Te–Ag core–shell nanowires, while Ag bulk diffusion in Te yields segmented Ag₂Te–Te nanowires. Electromigration is the primary driving force, and Joule heating dictates the diffusion pathway by modulating local temperature. Growth and transformation rates are tunable with applied bias and current density. This approach offers a controllable route for constructing complex, single-nanostructure heterostructures and provides mechanistic insight into diffusion-controlled nanoscale reactions. Future work could extend this strategy to other material systems, quantitatively map temperature and contact effects, and integrate the tailored heterostructures into functional nanodevices.

Quantitative voltage-dependent kinetics are difficult to extract due to variability and poor quantifiability of nanoscale contact conditions inside the TEM, including large and variable contact resistance. The formation process is limited by contact stability and geometry; thinner NWs react faster, complicating generalization across samples. Te nanowires may fracture near the phase transformation front due to Te’s low thermal evaporation temperature, affecting measurement windows. Temperature estimates rely on modeling and typical parameters rather than direct local thermometry. Minor thermal diffusion may occur concurrently, but electromigration dominates.