The 1000x efficiency headline is from Figure 4 of the source PDF: https://advances.sciencemag.org/content/advances/7/23/eabe4206.full.pdf https://advances.scie

The 1000x efficiency headline is from Figure 4 of the source PDF: https://advances.sciencemag.org/content/advances/7/23/eabe4206.full.pdf https://advances.sciencemag.org/content/advances/7/23/eabe42... Units are Amps per miliwatt on the Y axis and electron volts on the X axis. The black BTO line has a minimum of around 1x10^-10 at 2.8 eV, which is the energy of a photon in 450 nm blue light, and the red SBC222 line has a minimum of around 1x10^-7. Yes, the difference between 10^-10 and 10^-7 is 1000x; good job, headline! But no, that's not the key takeaway from the article. Solar cells have an efficiency curve based on their bandgap and the energy in the incident light. It's theoretically a straight line, and regular silicon cells get pretty close to it, see the figure here: https://www.pveducation.org/pvcdrom/solar-cell-operation/spectral-response https://www.pveducation.org/pvcdrom/solar-cell-operation/spe... Note the units in A/W, not A/mW, also, the research is at a temperature of 77k, the boiling point of liquid nitrogen, while the above plot is close to room temperature. At 450 nm = 2.a silicon cell produces about 0.35 x 10^-3 A/mW, which represents an efficiency of something like 30% in the real world. This research and that figure 4 plot fundamentally show that conventional barium titanate = BaTiO3 = BTO, produces very little energy from 450 nm light. The researchers fixed this problem with a multilayer cell using strontium titanate, SrTiO3, or STO, then BTO, then calcium titanate, CaTiO3 or CTO, in an STO_2/BTO_2/CTO_2 crystal lattice stack. That's the breakthrough result; they were able to bridge this gap in energy output and take in more wavelengths of visible light. The actual efficiency of the starting point was about a million times worse than a good silicon cell. There's plenty of room for improvement when you start that low! The end result is still on the order of a thousand times worse than a silicon cell. A production dye-sensitized solar cell today reaches about 10% efficiency, a monocrystaline silicon cell gets about 22%, but the DSSC has potential to be much cheaper to produce, can be flexible, and is not as well understood. This research may help to close that efficiency gap.