Since the end of the 20th century, numerous studies have analyzed Chinese economic development to gauge whether China's rapid growth is sustainable. Most of these studies focused on assessing total factor productivity (TFP) in Chinese mainland provinces but suffered from methodological weaknesses by assuming constant returns to scale (CRS) for the production frontier and/or incorrectly modeling variables returns to scale (VRS) technology taking into account bad output such as carbon dioxide emissions. Our paper offers a right non-parametric programming framework based on weak disposability and VRS assumptions to estimate environmental growth convergence among Chinese regions characterized by size heterogeneity. We explicitly separate regional efficiency gaps into two components: The first studies the technical catching-up process on each one (technical effect), and the second reveals convergence or divergence in the combinations of input and output among regions (structural effect). Moreover, carbon shadow price levels for provinces can be derived through the dual version of our activity analysis framework. Our empirical work focuses on 30 Chinese regions from 1997 to 2010. The results emphasize that environmental growth convergence among regions has mainly relied on the structural effect. We find that the structural effect largely depends on the pollution cost convergence and not on the evolution of the relative prices of capital or labor. The carbon shadow price is increasing at an annual rate of 2.5% and was evaluated around 864 yuan per ton in 2010 in China while regional estimates show significant disparities at the beginning of the period.