LEDs with test currents of at least 350 mA are considered. The following sub-efficiencies are represented: ηVf , forward voltage efficiency; ηIQ , internal quantum efficiency; ηdroop , efficiency droop; ηLE , light extraction efficiency; ηCR , conversion efficiency for red phosphors; ηCY/G , conversion efficiency for yellow/green phosphors; ηS , spectral efficiency. The overall LED lamp efficiency ηL is displayed in the right-most column. A horizontal line at 100% efficiency of an LED with a colour rendering index (CRI) of 85, and a colour temperature (CT) of 3,800 K is provided for visual guidance. Vertical bars represent cumulative contributions of LED technology innovations to efficiency improvements, with purple bars indicating specific innovations driven by technology spillovers and grey bars cumulatively indicating all other innovations identified in this study. Horizontal coloured lines indicate sub-efficiency levels of best-performing LEDs for the four years used in Fig. 2 : 2003, 2010, 2016 and 2020. For additional historical context, data for 1997 is included whenever possible. ‘NA’ denotes sub-efficiencies where 1997 data could not be calculated for the following reasons: ηVf, ηdroop depend on the device current, which was below 350 mA in 1997, making a comparison with contemporary devices difficult. ηCR, ηCY/G and ηS are relevant only to warm white LEDs, which were not available in 1997. ITO, indium tin oxide; PSS, patterned sapphire substrate; Phos., phosphors. Note that the ITO current spreading layer affects different sub-efficiencies in different chip architectures. For example, in modern flip-chip architectures ηLE no longer depends on ITO (Supplementary Note 2 ). Physical limits on sub-efficiencies are indicated by black horizontal lines. Note that the physical limit of ηVf is above 100% due to quantum effects 94 . Δ represents the remaining gap between the most recent values and physical limits for each sub-efficiency. Source: own elaboration based on data represented in Fig. 2 , Table 2 and Supplementary Note 3 .
