空気を掃除機で吸う：直接空気回収の残酷な物理学

The math of climate change is simple: humanity has emitted 2.5 trillion tons of CO2 since the Industrial Revolution. Even if every car switches to electric and every power plant converts to solar tomorrow, that blanket of carbon remains.

Nature’s solution, primarily trees, is elegant but slow. A mature tree captures about 22kg of CO2 per year. To offset global emissions, a forest roughly the size of Russia plus Canada combined would be required, and it would take 50 years to mature.

Enter Direct Air Capture (DAC). The premise is sci-fi: Giant industrial fans that vacuum the atmosphere, stripping out the carbon dioxide and burying it underground forever. It sounds like the perfect techno-fix.

But there is a problem. It relies on fighting the Second Law of Thermodynamics, which imposes a steep price for reversing entropy.

The Entropy Penalty: 420 PPM vs. 10%

Engineers have been scrubbing CO2 from gas streams for decades. In a coal power plant or a cement factory, the exhaust (flue gas) is about 10-15% CO2. In the open atmosphere, the concentration of CO2 is roughly 0.042% (420 parts per million).

This difference isn’t just a matter of “processing more air.” It creates a massive Entropy Penalty. Thermodynamics dictates that separating a dilute substance requires exponentially more energy than separating a concentrated one. The process is essentially un-mixing a gas, which reverses the natural state of disorder.

The minimum theoretical work (Gibbs Free Energy) to separate CO2 follows this logarithmic relationship:

$W_{min} = -RT \ln\left(\frac{P_{start}}{P_{end}}\right)$

• Flue Gas (10%): The CO2 is already concentrated. The partial pressure is high, meaning the gas molecules are easier to grab. The minimum thermodynamic work is remarkably low, roughly 0.1-0.2 GJ/tonne.
• Ambient Air (0.04%): The system is fighting the natural tendency of gases to mix (High Entropy). Finding a CO2 molecule in ambient air is statistically improbable compared to flue gas.

Because of this dilution, a DAC plant must process 2,500 times more air than a flue gas scrubber to capture the same amount of Carbon. This means giant fans, massive contactors, and a huge energy bill just to push the air through the filters. The pressure drop alone (the energy lost pushing air through a filter) becomes a dominant cost factor when moving gigatons of atmosphere.

In the real world, efficiency losses mean the energy cost is staggering:

• Theoretical Minimum: ~0.45 GJ/tonne (roughly 125 kWh/tonne).
• Real World (Current): ~8-11 GJ/tonne (2,200 - 3,000 kWh/tonne).

To put that in perspective: To capture 1 gigaton of CO2 (annual emissions are ~37 gigatons), the world would need roughly 2,500 TWh of clean electricity. That is equivalent to the entire annual electricity consumption of the United States.

The Tech Stack: Solids vs. Liquids

Despite the brutal physics, two main approaches have emerged to fight this battle. They are currently scaling up in the deserts of Texas and the lava fields of Iceland.

1. Solid Sorbent (The “Climeworks” Approach)

The Mechanism: This method uses a giant air filter, similar to a car’s intake, but coated with a specialized chemical (amine) that acts like chemical velcro for CO2. The amines chemically bond with CO2 molecules at ambient temperatures.

1. Capture Phase: Fans draw ambient air through the collector. The CO2 molecules adhere to the amine-coated filter material inside. The air passes through, stripped of its carbon.
2. Regeneration Phase: Once the filter is saturated (full), the collector closes. It is heated to roughly 80°C - 100°C. This low-grade heat breaks the weak chemical bond, releasing pure CO2 gas which is then collected.

• Thermodynamic Advantage: It runs at relatively low temperatures. This allows the system to run on waste heat from industrial processes or geothermal energy (as seen in Iceland’s “Orca” and “Mammoth” plants).
• Engineering Challenge: The physical filters degrade over time due to oxidation and thermal stress. The vacuum/heating cycle (Temperature Swing Adsorption, or TSA) is complex to manage at megaton scale, requiring thousands of modular “cubes” opening and closing in sequence.

2. Liquid Solvent (The “Occidental” Approach)

The Mechanism: This approach, pioneered by Carbon Engineering and now deployed by Occidental Petroleum (1PointFive), uses chemical engineering brute force. It relies on a liquid solution to wash the air.

1. Air Contactor: Huge fans pull air through a “honeycomb” packing structure where a liquid solution of Potassium Hydroxide (KOH) drips down. The CO2 reacts with the KOH to form Potassium Carbonate (salt).
2. Pellet Reactor: The salt solution is pumped into a reactor where it mixes with Calcium Hydroxide. The carbon transfers to the Calcium, forming small pellets of Calcium Carbonate (effectively limestone).
3. Calciner: These pellets are fed into a massive kiln heated to 900°C. The extreme heat calcines the limestone, releasing pure CO2 and leaving behind Calcium Oxide, which is recycled back into the process.

• Thermodynamic Advantage: It uses standard industrial chemicals (KOH, Calcium) available by the ton. The process can run continuously with no “batch” cycling, making it easier to scale to massive sizes (like the Stratos plant in Texas).
• Engineering Challenge: The thermodynamic penalty is high. Reaching 900°C requires massive amounts of energy. Originally, this was designed to run on natural gas, with the plant capturing its own emissions. Transitioning this high-heat step to electric or hydrogen power is a significant hurdle.

The Water Equation

Thermodynamics isn’t the only limit; geography plays a role. Liquid solvent systems rely on aqueous solutions. In dry environments (like the Permian Basin where Stratos is being built), water evaporation can be a major issue. While designs optimize for water recovery, “vacuuming” millions of cubic meters of dry desert air inevitably strips moisture from the solvent. Solid sorbents are generally more water-neutral (or can even capture water as a byproduct), giving them a geographic edge in arid climates, though they suffer more from dust and particulate clogging.

The Cost of Purity

The “Energy Penalty” translates directly into a “Cost Penalty.” Currently, capturing a ton of CO2 via DAC costs between $600 and $1,000. The United States Department of Energy has set a “Earthshot” goal of $100/ton by 2032.

Is that possible? Thermodynamics says yes, but engineering says it will be a fight. The theoretical minimum (125 kWh/ton) suggests there is large headroom for efficiency gains. However, the industry is fighting the law of diminishing returns. As the concentration of CO2 in the air rises (a disaster for the climate), DAC actually becomes slightly more efficient. But relying on catastrophic pollution levels to improve machine efficiency is a grim optimization strategy.

The path to $100/ton requires:

1. Sorbent Longevity: Filters that last years, not months.
2. Passive Airflow: Removing the fans (using natural wind) to cut the electric load.
3. Heat Integration: Colocating with nuclear SMRs or geothermal sources to get “free” thermal energy.

The Verdict

DAC is not a replacement for cutting emissions. The thermodynamics are too punishing. It makes no fiscal or physical sense to burn coal to power a DAC plant to clean up the coal’s mess. The energy return on investment (EROI) is negative.

DAC only makes sense when powered by stranded renewables. For example, using solar power at noon in the desert, or geothermal in remote locations that cannot reach the grid. In these specific edge cases, it transforms excess biological energy into a geological service.

It is the vacuum cleaner of last resort. Humanity shouldn’t dump trash on the floor just because a vacuum exists, but considering the 2.5 trillion tons of mess already accumulating, a very big machine is likely necessary to clean it up.