Extended Data Fig. 1: Lack of mass-dependent isotope fractionation in the step-crushing neon data.

Péron et al.⁸ suggested that using the highest measured ²⁰Ne/²²Ne ratios for characterizing the plume mantle is inappropriate because mass-dependent isotope fractionation may occur during bubble formation⁸. In this scenario, mass-dependent fractionation during bubble formation would lead to ²⁰Ne/²²Ne ratios scattering about the ‘true’ mean value, with some bubbles characterized by relatively high ²⁰Ne/²²Ne ratios while other bubbles displayed relatively low ²⁰Ne/²²Ne ratios. This hypothesis can be tested by measuring ⁴He/³He and ³⁸Ar/³⁶Ar ratios during the same step-crushes as the neon isotopes as illustrated here. a, b, The measured neon isotopic compositions of individual step-crushes plotted against those of helium (a) and argon (b) along with predicted trajectories of mass-dependent isotope fractionation during bubble formation obtained by applying a Rayleigh fractionation model⁷⁸. Here, the parental melt is assumed to have an initial ²⁰Ne/²²Ne ratio of 12.65 ± 0.08, similar to the value of ref. ⁸. Initial helium and argon isotopic compositions are from the mean values determined for sample EW9309_5D in this study (Supplementary Table 1). The curves show the trajectory of the melts during degassing, the evolution of an instantaneously lost vapour phase (bubbles; short-dashed line) and the cumulative evolution of the vapour phase (bubbles; solid line). Plotted along with these curves are the individual step-crushes (circles) from this study (sample EW9309_5D) with their associated 2σ uncertainties (error bars). Note that the ⁴He/³He ratios were measured only on one aliquot of EW9309_5D. The helium–neon–argon isotopic compositions measured in the individual step-crushes do not follow the predicted Rayleigh fractionation trends. Rather, the data cloud is at a high angle to the predicted isotope fractionation trend. For example, for our highest measured ²⁰Ne/²²Ne of 13.03 ± 0.04 (2σ) to be a result of mass fractionation, the measured ³⁸Ar/³⁶Ar should be 0.1847 (b). However, the measured ³⁸Ar/³⁶Ar of 0.1885 ± 0.0014 (2σ) is identical to the atmospheric value and similar to other determinations of ³⁸Ar/³⁶Ar ratios in plumes and MORBs. Given this, we conclude that mass-dependent isotope fractionation during bubble formation is not responsible for generating the highest ²⁰Ne/²²Ne ratios determined in these studies.