H ii regions are ionized nebulae surrounding massive stars. They exhibit a wealth of emission lines that form the basis for estimation of chemical composition. Heavy elements regulate the cooling of interstellar gas, and are essential to the understanding of several phenomena such as nucleosynthesis, star formation and chemical evolution^1,2. For over 80 years³, however, a discrepancy exists of a factor of around two between heavy-element abundances derived from collisionally excited lines and those from the weaker recombination lines, which has thrown our absolute abundance determinations into doubt^4,5. Here we report observational evidence that there are temperature inhomogeneities within the gas, quantified by t² (ref. ⁶). These inhomogeneities affect only highly ionized gas and cause the abundance discrepancy problem. Metallicity determinations based on collisionally excited lines must be revised because these may be severely underestimated, especially in regions of lower metallicity such as those recently observed with the James Webb Space Telescope in high-z galaxies^7,8,9. We present new empirical relations for estimation of temperature and metallicity, critical for a robust interpretation of the chemical composition of the Universe over cosmic time.

H ii区是大质量恒星周围的电离星云。它们展示出大量的发射线，这些发射线构成了估计化学成分的基础。重元素调节星际气体的冷却，对于理解核合成、恒星形成和化学演化等多种现象至关重要^1,2。然而， 80 多年来³，碰撞激发线的重元素丰度与较弱的复合线的重元素丰度之间存在大约两倍的差异，这使我们对绝对丰度的确定产生了怀疑^4,5。在这里，我们报告了气体内存在温度不均匀性的观测证据，用t量化² （参考文献⁶）。这些不均匀性仅影响高度电离的气体并导致丰度差异问题。必须修改基于碰撞激发线的金属丰度测定，因为这些可能被严重低估，特别是在金属丰度较低的区域，例如最近使用詹姆斯韦伯太空望远镜在高 z 星系中观察到的区域^7,8,9。我们提出了用于估计温度和金属丰度的新经验关系，这对于可靠地解释宇宙在宇宙时间内的化学成分至关重要。