Extended Data Figure 1: Comparisons of ELVOC spectra.

a, b, Comparison of CI-APi-TOF spectra from Hyytiälä and JPAC, during night (a) and day (b). The JPAC spectra (light red) show ELVOCs clustered with ¹⁵NO₃⁻ whereas Hyytiälä data are clustered with ¹⁴NO₃⁻, and therefore identical ELVOCs will be found shifted by 1 Th. The features of the night-time spectrum measured in Hyytiälä are very similar to those in the JPAC α-pinene ozonolysis (2 p.p.b. α-pinene, 17 p.p.b. ozone) spectrum (a), with the same pattern of peaks, as well as most major single peaks in good agreement. The day-time Hyytiälä ELVOC spectrum is also replicated quite well, but here using conditions where α-pinene is mainly oxidized by OH, in the presence of NO_x (0.7 p.p.b. α-pinene, 20 p.p.b. ozone, 10⁷ cm⁻³ OH, 30 p.p.b. NO_x, ∼3 p.p.b. NO). The patterns overlap, and although some single peaks are only found in one of the spectra, the major peaks such as the organic nitrates at 339 Th and 371 Th (C₁₀H₁₅NO₉ and C₁₀H₁₅NO₁₁) in the Hyytiälä spectrum are found at 340 Th and 372 Th in the JPAC spectrum, as expected. The fraction of α-pinene reacting with ozone in b is probably higher in Hyytiälä than under our chamber conditions. The use of labelled nitrate enables us to assign the N-atom to the ELVOCs and not to clustered HNO₃. The peak at 402 Th in the Hyytiälä spectrum in b, on the other hand, corresponds to the same ELVOC as 339 Th, with one additional HNO₃ adduct. The corresponding peak in the JPAC spectrum is now shifted to 404 Th, due to the presence of two ¹⁵N-atoms. c, d, ELVOC peak shifts due to isotopically labelled ozone during α-pinene ozonolysis measurements in the Tropos flow tube at 25% relative humidity, using 80 p.p.b. α-pinene. Ozone concentrations, both labelled and non-labelled, were roughly 15 p.p.b. c.p.s., counts per second.