> Air compression is not cheap, energetically. Chilling is also energetically expensive. You'll need to do both. Not exactly. The idea is that you don't just

> Air compression is not cheap, energetically. Chilling is also energetically expensive. You'll need to do both. Not exactly. The idea is that you don't just compress air - and let the heated compressed air cool to the atmosphere in order to then expand it so it would cool some more, and then just release whatever's left back to the atmosphere. No, you use those leftovers to cool the incoming air - so you need to compress air spending less energy - and you also get the expanding air to help you to compress incoming air. The principal point where you have non-reversibility is that extra cooling - making compressed cold air to do the work - which you need to reach necessary temperature. And you only need that temperature to separate gases; after that you can use the cold products to help with compression. So - without colorful metaphors - I'd like to figure out how much energy will you need. I see the approach of SES with flue gases. I'm not sure how it will look like with regular air.