Direct air capture: process technology, techno-economic and socio-political challenges

This is a **systematic review** (not an original experiment) that synthesises findings from hundreds of published studies on direct air capture. The authors examined:

**Technologies tested:** Solid sorbent DAC (using materials like amine-functionalised silica, metal-organic frameworks, and zeolites) and liquid solvent DAC (using aqueous hydroxide solutions, typically potassium hydroxide or sodium hydroxide). They also reviewed emerging approaches like electrochemical DAC and membrane-based capture.

**Comparators:** No direct comparator — this is a state-of-the-field review. However, they implicitly compare DAC to other negative emissions technologies (e.g., bioenergy with carbon capture and storage, afforestation) and to point-source carbon capture (capturing CO₂ from industrial exhaust stacks).

**Outcome measures:** CO₂ capture capacity (mol CO₂ per kg sorbent), energy requirement (GJ per tonne CO₂), cost ($ per tonne CO₂), energy penalty (fraction of energy output consumed by capture), water consumption (kg water per tonne CO₂), and lifecycle emissions (kg CO₂-equivalent per tonne CO₂ captured).

No human participants were studied. This is a review of **laboratory-scale experiments, pilot plants, and techno-economic models**. The studies reviewed span:

**Laboratory studies:** Typically using synthetic air (400–500 ppm CO₂) in small-scale reactors (grams to kilograms of sorbent), with controlled temperature (25–120°C) and humidity (0–100% relative humidity).

**Pilot plants:** The most well-known is Climeworks' plant in Hinwil, Switzerland (operational since 2017, capturing ~900 tonnes CO₂ per year using solid sorbents). Another is Carbon Engineering's pilot in Squamish, Canada (capturing ~1 tonne CO₂ per day using liquid solvents). A third is Global Thermostat's pilot in Alabama, USA.

**Techno-economic models:** These are computer simulations that estimate costs for hypothetical large-scale plants (capturing 0.1–1 million tonnes CO₂ per year), assuming various energy sources (natural gas, waste heat, renewables), locations (deserts, coastal areas, industrial sites), and financing scenarios.

The review compiled data from published studies using standardised metrics:

**CO₂ capture capacity:** Measured via thermogravimetric analysis (TGA) or breakthrough curves in packed-bed reactors. Units: mmol CO₂ per g sorbent.

**Energy requirement:** Calculated from the heat needed for sorbent regeneration (typically 80–120°C for solid sorbents, 800–900°C for liquid solvents) plus electricity for fans, pumps, and compressors. Units: GJ per tonne CO₂.

**Cost:** Derived from techno-economic models using discounted cash flow analysis, capital expenditure (CAPEX), and operating expenditure (OPEX). Units: $ per tonne CO₂ captured (or $ per tonne CO₂ avoided, which accounts for emissions from the energy source).

**Lifecycle emissions:** Calculated using lifecycle assessment (LCA) methodology, accounting for emissions from sorbent production, plant construction, energy use, and sorbent disposal. Units: kg CO₂-equivalent per tonne CO₂ captured.

**Water consumption:** Measured as net water loss from the system (evaporation, cooling tower blowdown, and water used in sorbent regeneration). Units: kg water per tonne CO₂.

### Study design

This is a **comprehensive narrative review** with systematic elements. The authors searched multiple databases (Web of Science, Scopus, Google Scholar) for peer-reviewed articles, conference proceedings, and industry reports published up to 2021. They did not perform a formal meta-analysis (no pooled effect sizes or forest plots) but instead synthesised findings qualitatively and presented ranges of reported values.

### How they selected studies

The authors state they "appraised the state-of-the-art" but do not provide a PRISMA diagram or explicit inclusion/exclusion criteria. This is a major methodological weakness — it means the review is vulnerable to selection bias (the authors may have preferentially included studies that support their narrative).

### What the design can and cannot prove

**What it can prove:**

The range of reported performance metrics across different DAC technologies.

The current technical bottlenecks (e.g., sorbent degradation, energy requirements).

The consensus (or lack thereof) on cost estimates.

The policy and social challenges identified in the literature.

**What it cannot prove:**

Which technology is "best" — because studies use different assumptions, scales, and conditions, direct comparisons are unreliable.

The true cost of DAC at scale — because no commercial-scale plant (capturing >1 million tonnes CO₂ per year) has been built yet. All cost estimates are extrapolations from pilot plants and models.

Long-term sorbent stability — because most laboratory studies run for <100 cycles (days to weeks), while commercial plants would need >10,000 cycles over decades.

Real-world performance under variable ambient conditions — because most studies use controlled laboratory conditions (constant temperature, humidity, CO₂ concentration).

### Major methodological weaknesses

1. **No systematic search protocol** — cannot be replicated.

2. **No quantitative synthesis** — cannot assess heterogeneity or publication bias.

3. **Industry-funded studies** — many techno-economic models are funded by companies developing DAC (e.g., Carbon Engineering, Climeworks), creating potential conflicts of interest.

4. **Outdated cost estimates** — the review includes studies from 2010–2021, but DAC costs have changed rapidly. For example, Climeworks announced costs of $600–$800/tonne in 2021, but their 2022 Mammoth plant claims costs of ~$300/tonne.

5. **Geographic bias** — most studies assume deployment in North America or Europe, with little analysis of developing country contexts.

### Technical performance

**Solid sorbent DAC:** Reported CO₂ capture capacities range from **0.5–3.0 mmol CO₂ per g sorbent** (typical: ~1.5 mmol/g for amine-functionalised silica). Energy requirements: **5–10 GJ per tonne CO₂** for regeneration (heat at 80–120°C). Sorbent degradation: **5–30% capacity loss after 100 cycles** due to amine oxidation and urea formation.

**Liquid solvent DAC:** Reported capture capacities: **0.5–1.0 mol CO₂ per L of solution** (for potassium hydroxide). Energy requirements: **8–15 GJ per tonne CO₂** for regeneration (heat at 800–900°C). Water consumption: **1–7 tonnes water per tonne CO₂** (due to evaporative losses in the calciner).

**Emerging technologies:** Electrochemical DAC (using pH swings or redox reactions) reports energy requirements of **2–5 GJ per tonne CO₂** but only at laboratory scale (<1 g CO₂ per day). Membrane-based DAC reports **0.1–0.5 mmol CO₂ per m² per second** but requires high pressure differentials.

### Cost estimates

**Current costs (2021):** **$100–$1,000 per tonne CO₂** captured. The wide range reflects different assumptions about energy costs, plant scale, and financing.

**Future cost projections (2030–2050):** **$50–$200 per tonne CO₂** if learning rates follow historical patterns for similar technologies (e.g., solar PV, wind). However, the authors note that DAC has not yet demonstrated any learning-by-doing because no commercial plants exist.

**Breakdown:** Energy costs account for **40–60%** of total cost. Capital costs account for **30–50%**. Sorbent replacement accounts for **5–15%**.

**Cost floor:** The authors estimate a thermodynamic minimum cost of **~$20 per tonne CO₂** (based on the Gibbs free energy of mixing CO₂ from 400 ppm to pure CO₂), but real-world costs are 5–50× higher due to inefficiencies.

### Energy and environmental impacts

**Energy penalty:** DAC requires **5–15 GJ per tonne CO₂** of thermal energy plus **0.5–2 GJ per tonne CO₂** of electricity. For context, capturing 1 billion tonnes CO₂ per year (the scale needed for climate targets) would require **5–15 EJ per year** — equivalent to **10–30% of current global electricity generation**.

**Lifecycle emissions:** If powered by natural gas (without carbon capture), DAC has lifecycle emissions of **0.1–0.3 tonnes CO₂-equivalent per tonne CO₂ captured** (i.e., it is 70–90% net negative). If powered by renewables, lifecycle emissions drop to **0.02–0.05 tonnes CO₂-equivalent per tonne CO₂ captured** (95–98% net negative).

**Water consumption:** Liquid solvent DAC consumes **1–7 tonnes water per tonne CO₂** (mostly from evaporative cooling). Solid sorbent DAC consumes **0.1–0.5 tonnes water per tonne CO₂** (for sorbent regeneration and cooling). For context, capturing 1 billion tonnes CO₂ per year with liquid DAC would consume **1–7 billion tonnes of water** — equivalent to the annual water use of **200–1,400 million people** (at 50 L/person/day).

### Policy and social challenges

**Carbon pricing:** Most models assume a carbon price of **$100–$200 per tonne CO₂** to make DAC economically viable. Current carbon prices are much lower: EU ETS ~$50/tonne, California ~$30/tonne, and most of the world has no carbon price.

**Public acceptance:** Surveys show **40–60% support** for DAC in the US and Europe, but support drops to **20–30%** when respondents are told about costs or energy requirements. There is also concern about "moral hazard" — the risk that DAC reduces urgency for emissions reductions.

**Governance:** No international framework exists for accounting, verifying, or trading DAC-based carbon removal. The authors note that "without clear rules, DAC credits could be double-counted or used to offset emissions that never occurred."

Since this is a review, there is no single "effect size." Instead, the key magnitudes are:

**Cost gap:** The current cost of DAC ($100–$1,000/tonne) is **10–100× higher** than the cost of avoiding emissions (e.g., energy efficiency at $10–$50/tonne, solar PV at $20–$50/tonne). This means DAC is currently one of the most expensive climate mitigation options.

**Scale gap:** To remove 1 billion tonnes CO₂ per year (roughly 2–3% of global emissions), you would need **1,000–10,000 DAC plants** each the size of the largest planned facility (Mammoth, capturing ~36,000 tonnes/year). Building that many plants would require **$100 billion–$1 trillion** in capital investment.

**Energy gap:** The energy required for DAC at climate-relevant scales (5–15 EJ/year) is equivalent to **10–30% of current global electricity generation** — meaning DAC would compete directly with other energy uses (homes, industry, transport) unless powered by dedicated renewables.

**Time gap:** Even with aggressive deployment, the authors estimate it would take **10–20 years** to build enough DAC capacity to remove 1 billion tonnes CO₂ per year. Given that we need to reach net-zero by 2050, this means DAC cannot be a primary solution — it can only be a supplementary tool.

### What the authors acknowledge

1. **Data scarcity:** "There is a lack of long-term, real-world performance data for DAC systems operating under variable ambient conditions."

2. **Model uncertainty:** "Techno-economic models rely on assumptions that may not hold at scale, particularly regarding sorbent lifetime, energy costs, and financing."

3. **Publication bias:** "Studies reporting positive results (high capture capacity, low cost) are more likely to be published than those reporting negative results."

4. **Geographic limitations:** "Most studies assume deployment in locations with low-cost energy and water, which may not be available in many regions."

### What a critical reader would note

1. **No systematic search protocol** — the review cannot be replicated, and the authors may have selectively included studies.

2. **Industry funding** — many of the cited techno-economic models were funded by DAC companies (Carbon Engineering, Climeworks), which have a financial interest in showing low costs.

3. **Outdated data** — the review includes studies from 2010–2021, but DAC technology is evolving rapidly. For example, the cost estimates do not include recent advances in electrochemical DAC or novel sorbents.

4. **No comparison to alternatives** — the review does not systematically compare DAC to other negative emissions technologies (e.g., bioenergy with CCS, enhanced weathering, afforestation), making it hard to assess relative merits.

5. **No sensitivity analysis** — the review presents cost ranges but does not show how sensitive costs are to key assumptions (e.g., sorbent lifetime, energy price, discount rate).

6. **Narrow focus** — the review focuses on technical and economic aspects but gives limited attention to social justice, land use, or ecological impacts of large-scale DAC deployment.

For someone running their own n=1 experiment (e.g., testing a small-scale DAC device or evaluating carbon removal claims):

### What to test

**Specific intervention:** If you have access to a small DAC device (e.g., a desktop unit from a company like Climeworks or a DIY prototype), test its CO₂ capture rate under different ambient conditions. Alternatively, test the performance of a specific sorbent material (e.g., amine-functionalised silica, zeolite 13X, or metal-organic framework) in a small packed-bed reactor.

**Dose:** For a sorbent test, use a fixed mass (e.g., 10 g) and expose it to air with controlled CO₂ concentration (400–500 ppm) at a fixed flow rate (e.g., 1 L/min). For a device test, run it at its maximum rated capacity.

### Minimum meaningful duration

**Sorbent cycling:** Run at least **50–100 adsorption-desorption cycles** (each cycle: 30–60 minutes adsorption, 10–30 minutes desorption). This is the minimum to assess sorbent degradation. Most laboratory studies run 100 cycles, but commercial plants need >10,000 cycles.

**Device testing:** Run for **at least 7 days** of continuous operation (24/7) to capture diurnal variations in temperature, humidity, and CO₂ concentration. Longer is better — 30 days would give more reliable data.

### What to measure (specific metrics)

**CO₂ capture capacity:** Measure the mass of CO₂ captured per cycle (using a mass balance or CO₂ sensor) and divide by sorbent mass. Units: mmol CO₂ per g sorbent. Target: >1.0 mmol/g for a decent sorbent.

**Energy consumption:** Measure the electricity used by fans, pumps, and heaters (using a power meter). Also measure the heat input for regeneration (using a temperature sensor and flow meter). Units: kWh per kg CO₂ captured. Target: <10 kWh/kg CO₂ (equivalent to ~36 GJ/tonne CO₂) for a decent system.

**Sorbent degradation:** Measure the capture capacity after every 10 cycles. Plot capacity vs. cycle number. A linear fit gives the degradation rate (e.g., 0.1 mmol/g per 100 cycles). Target: <10% loss after 100 cycles.

**Water consumption:** If using a liquid solvent, measure the volume of water added to maintain solution concentration. Units: L water per kg CO₂ captured. Target: <1 L/kg CO₂ for solid sorbents, <5 L/kg CO₂ for liquid solvents.

**Ambient conditions:** Record temperature (°C),