The window of opportunity for effective climate change mitigation is closing. Hence, it is decisive to understand how to accelerate the diffusion of climate friendly technologies. Path dependence of technological change is an explanation for sluggish diffusion even if a technology is superior in the long run. This paper studies the determinants of diffusion, learning and the coevolution of innovation and heterogeneous absorptive capacity. I show how the effectiveness of different market based climate policies depends on the type and strength of diffusion barriers. I introduce a macroeconomic agent-based model that is an eco-technology extended version Eurace@unibi model. Technology is heterogeneous by type (green or brown). Firms choose between types when acquiring capital goods and build up type-specific technological know-how that is needed to exploit the productive potential of capital. Path dependence is operationalized as accumulated diffusion barriers taking the form of inferior technical performance of supplied green capital and type-specific know-how of adopters. The barriers interrelate with positive feedback loops from market induced innovation dynamics and learning by doing, and analyze how these mechanisms explain path dependence and the emerging macroeconomic patterns of technology diffusion. Environmental taxes can outweigh a lower technical performance and subsidies perform better if lacking capabilities hinder firms to adopt a sufficiently mature technology.