Fig. 4: WS₂ films as a diffusion barrier.

a, Schematic of the Cu/SiO_2|nat/Si structures used for the thermal stress tests. b,c, Elemental distribution maps of Cu/SiO_2|nat/Si after the thermal stress on a plain Si substrate (b) and on a substrate with trench structure (c), with EDS elemental mapping of Cu as inset. d, Schematic of the Cu/single-layer WS₂/SiO_2|nat/Si structures used for assessing WS₂ film as diffusion barriers under thermal stress. e,f, Elemental distribution maps of Cu/single-layer WS₂/SiO_2|nat/Si after the thermal stress on a plain Si substrate (e) and on a substrate with trench structure (f), with EDS elemental mapping of Cu as inset. g–i, GIXRD analysis before and after the thermal stress. g, Reference Cu and Cu₃Si peaks (Methods). h,i, GIXRD spectra for the Cu/SiO_2|nat/Si (h) and Cu/single-layer WS₂/SiO_2|nat/Si (i), both before and after annealing. j, Device structure used for electrical breakdown measurements to assess WS₂ films as diffusion barriers under electrostatic stress. k, Measured characteristic breakdown time τ as a function of applied voltage for single-layer WS₂ and a reference without a barrier. Dashed lines show extracted E model projections for the reference (orange) and a single-layer WS₂ barrier (green). Each τ value was extracted from 5–10 breakdown-time measurements in different junctions. Error bars show the standard deviation of log₁₀ [τ (s)]. Individual breakdown-time data points are displayed in Supplementary Section 10. l, Cumulative distribution of the projected breakdown time for an electrical stress of 8 MV cm⁻¹ from the measurements. as-dep., as deposited; BF, bright field; ref., reference.