As with any difficult problem, one should start from first principles. In the case of atmospheric carbon these turn out to be its three main physical characteristics:

1. Volume
2. Sparseness
3. Mass

Volume. Earth’s atmosphere stretches from planet’s surface up to as far as 10,000 kilometers, with total volume of some 5.2*10¹⁹ m³. This is over four times the volume of world’s oceans. Whilst carbon dioxide is a relatively heavy molecule and thus denser in lower altitudes, the necessity to shift through massive air volumes requires exposing very large surface areas. Volume wise you are boiling the ocean – many times over.

Sparseness. Contra to the large total volume, the number of carbon molecules per unit of air is actually small. Whilst double from the start of the industrial revolution, the current concentration is still only 421 particles per million, i.e. very sparse. Capturing CO₂ molecules from air is like looking for a needle in a haystack. Therefore filtering through the particles must be very energy efficient – you capture only four needles per every ten thousand examined.

Mass. Final key characteristic is heavy mass. One ton is already a lot – gigaton 10⁹ truly so. We emit annually about 37 gigatons of carbon dioxide. Therefore, upon finding and capturing the atmospheric carbon, the process must provide a very large mass for absorption and storage.

These three requirements – large surfaces, energy efficiency, and significant mass – are the main constraints for any carbon capture mechanism. These are also the reasons why it appears unlikely that any of the currently developed capture technologies will make the necessary cost-effective, large-scale impact. At least that was our assessment from studying them quite extensively.

Indeed, there turns out to be only three possible technologies with the required surfaces, efficiency, and mass:

1. Oceans
2. Landsurfaces
3. Forests