3. Results
Using today's technology, electrofuel production using commercially available low-temperature DAC would be substantially more expensive than petroleum jet fuel, shown in Figure 2 at $4.41–$4.66 per liter of gasoline equivalent (lge), compared to the average US pre-tax jet fuel price of $0.59/lge from 2010 to 2019 (all in 2017$). However, costs in the next-decade and by 2050 cases, described in Materials and Methods, are dramatically lower, $1.68–1.76/lge and $0.95–0.99/lge, respectively, for high-temperature DAC, with almost identical costs for low-temperature DAC cases.
Figure 2. Levelized cost of electrofuel per liter of gasoline equivalent (lge) of liquid hydrocarbon fuel produced using solar (A) and wind (B) electricity, considering both high- and low-temperature DAC systems in the next-decade and by 2050 cases. In all cases, the three largest cost components are capital and non-energy operational costs for the DAC system, the electrolyzer, and the renewable electricity generation. Most cost components represent combined capital and non-energy operational costs, with the exceptions of natural gas, grid electricity, and water which represent the cost of commodities purchased from an external interconnection. The black dotted and dashed lines are the minimum, maximum, and average petroleum jet fuel prices from 2010 to 2019. All costs are in 2017$.
Note that cost parameters for these five scenarios do not attempt to represent a full range of possible futures in the development of key component technologies but instead illustrate possibilities based on the available literature. The relatively conservative today case, which relies largely on the peer-reviewed literature and government reports, may not capture cutting-edge developments. See Materials and Methods for further discussion of the scenarios. The sensitivity analysis explores the relative importance of each individual factor.
In all cases, the combined capital and non-energy operating costs of renewable electricity generation, the electrolyzer, and the DAC system are dominant, representing 62.3–85.7% of fuel cost in all cases. Renewable electricity generation is 24.6–38.9% of fuel cost in all cases, with the electrolyzer at 17.7–23.4% in wind cases and 27.5–36.5% in solar cases. DAC represents 16.4–23.2% of fuel cost, with a roughly equal cost share in the comparable wind and solar cases.
In the today case for both wind and solar, the optimization supplements on-site renewable electricity with substantial amounts of natural gas and grid electricity despite incurring a carbon price of $600/t(CO2) to offset the resulting emissions with offsite DACS to ensure carbon neutrality (excluding facility construction life-cycle emissions). In the today wind case, these carbon price payments constitute the third largest cost component, at 21.3% of fuel cost, falling to 16.2% in the corresponding solar case. This is in large part because the Fischer–Tropsch system modeled for fuel synthesis in the today case cannot operate flexibly, forcing the system to ensure full operation for the entire year, largely through carbon-intensive grid electricity and natural gas to balance strongly seasonal wind and moderately seasonal solar. See the Supporting Information, Section S5.3 for cases with no carbon price.
Costs in the next-decade cases remain in the range of $1.68/lge–$1.77/lge, still close to three times the ten-year mean price of pre-tax jet fuel, $0.59/lge, but in a similar cost range to comparable biobased jet fuel alternatives. 73.3–78.6% of fuel cost in these cases is capital and non-energy operational costs of the renewable electricity generation, the electrolyzer, and the DAC system. Fuel synthesis represents 9.6%–10.1% of fuel cost, while the cost of offsetting emissions from grid electricity and fossil natural gas is 5.4–7.7%, above the cost of the electricity and gas itself (including interconnection costs) at 3.5–4.2%.
In the by 2050 cases, costs fall to $0.95–0.99/lge, roughly 50% above 10-year mean jet fuel prices and only 15–22% above average jet fuel prices in 2012. In the high-temperature DAC cases, renewable electricity generation, electrolyzer, and DAC system costs represent 76–80% of fuel costs, rising to 83–86% in low-temperature DAC cases. This is due largely to the assumed ability to recycle heat from both the fuel synthesis and the electrolyzer to power the roughly 100 °C DAC amine regeneration process. The high-temperature systems continue to supplement their electric kiln with heat from fossil natural gas, amounting to 1.2-2.3% of the total cost plus an additional 3.3–6.3% to offset the emissions with a carbon price at the prevailing cost of offsite DAC ($130/t(CO2) in the high-temperature case). The low-temperature system can use recycled waste heat or heat from the electric kiln, which is less expensive to store at lower temperatures, to meet much of its heat demand, although it still uses some natural gas at 0.5–0.6% of the total cost plus 1.1–1.2% to offset the resulting emissions at $100/t(CO2), the assumed cost of standalone offsite low-temperature DACS. Thus, waste heat recycling allows low-temperature DAC to compete with lower capital cost high-temperature DAC.
All cases make extensive use of the system flexibility offered by various forms of energy and material storage as well as external fossil energy interconnections. In the absence of fossil energy interconnections, the today system produces 12% and 27% more expensive fuel for solar and wind systems, respectively, with greater reliance on costly hydrogen, electricity, and heat storage as well as overbuild and curtailment of renewable electricity generation capacity. Under next-decade assumptions, a no-fossil case yields smaller increases of 0.7–0.8% in solar cases and 7–11% for wind cases. By 2050 cases with no fossil carbon are at cost parity with cases that have fossil interconnections for solar and are 2–6% more expensive for wind cases. See the Supporting Information, Section S5.4 for further detail.
Another important and related form of flexibility is undersizing, also discussed in ref. Undersized components will operate at a higher capacity factor (total utilization as a fraction of maximum possible utilization) than the electricity generation source, as shown in Table 3. The most consistently undersized component (relative to renewable electricity generation capacity) is the capital-intensive DAC system, which operates above 90% utilization (total annual output as a fraction of potential annual output) in all solar cases and above 81% utilization in all wind cases. The fuel synthesis system is similar, with utilization above 84% in all solar cases and above 71% in all wind cases. Note that although the solar electric capacity factor is 29.3%, roughly half that of the wind capacity factor of 57.5%, the utilization of DAC and fuel synthesis is higher in the solar case due to less extreme seasonality in electricity availability (see the Supporting Information, Section S5.2 for further detail). In the solar cases, electrolyzers are also undersized to achieve a 38.8–42.1% utilization, whereas relatively inexpensive electric kilns are oversized in all cases, with utilization levels of 6.2–20.0% to transform excess electricity to heat for the DAC system as needed. This component undersizing provides substantial cost reductions, illustrated by a cost increase of 68% in the solar today case if all system components are required to (unphysically) operate at the same capacity factor as the renewable energy. Thus, although the components that provide system flexibility that enable undersizing (storage of hydrogen, heat, electricity, and CO2, and grid electricity and natural gas consumption including fossil emissions offset payments) constitute 25.6% of fuel cost, they bring savings on the order of twice their cost. See the Supporting Information, Section S5.5 for further detail.
Note that the DAC, fuel synthesis, and, to a lesser extent, the electrolyzer have a higher capacity factor than the renewable electricity in all cases.
Note that in the today and by 2050 low-temperature cases, the optimization dispenses entirely with the electric kiln, relying only on natural gas in the former, heavily supplemented by waste heat recycled from the fuel synthesis system and electrolyzer in the latter. See the Supporting Information, Section S5.1 for component capacity in each scenario. In addition, the model never builds liquid fuel storage as there are few if any benefits of undersizing the fuel export pipeline relative to the fuel synthesis plant. The solar next-decade cases and all wind cases dispense entirely with electricity storage, opting to simply use expensive grid electricity as a backup. Similarly, the today and next-decade cases, as well as the by 2050 high-temperature wind case, have no or almost no heat storage, opting for grid natural gas with a steep carbon price instead. All next-decade and by 2050 cases build CO2 storage for seasonal shifting. See the Supporting Information, Section S5.7 for further detail.
A final important form of flexibility is curtailment. 4–26% of all renewable electricity generated in each scenario is curtailed as waste. The ability to curtail allows the system to undersize components, avoiding the construction of expensive production or storage capacity that would rarely be utilized. See the Supporting Information, Section S5.1 for further detail.
For low-temperature cases, water produced from the DAC system is sufficient to supply the electrolyzer. However, water costs in high-temperature cases represent at most $0.01/lge or 1.4% of the total cost.
3.1. GHG Mitigation Cost-Effectiveness
The implicit GHG mitigation cost in these scenarios, compared with fossil jet fuel at even its highest annual average value since 2010 of $0.81/lge, is $1,312-$1,400/t(CO2) today, $315–350/t(CO2) in the next-decade, and $51–65/t(CO2) by 2050. This uses the method described in Section S5.8, which assumes that electrofuels have zero net carbon intensity by paying a carbon price for emissions offsets. If fossil jet fuel prices fall to their lowest annual average value since 2010 of $0.35/lge, these implicit mitigation costs each rise by $167/t(CO2). Thus, simply offsetting fossil jet fuel emissions with the DACS technology used in each scenario (costs shown in Table S7) is a more cost-effective net-zero option in all scenarios except the by 2050 cases with a high fossil jet fuel price. Sequestration costs, assumed to be negligible in the main model, would have to rise above $85/t(CO2), at the very high end of the existing literature, to make electrofuels cost-effective in any of the today or next-decade scenarios. See the Supporting Information, Section S5.8 for further detail.
3.2. Sensitivity: The Importance of Capital Cost and Flexibility
Figure 3 shows a sensitivity analysis that begins with the levelized cost of electrofuel in the next-decade solar high-temperature case, $1.68/lge, and varies each parameter to its today and by 2050 values, with a few exceptions described in the Supporting Information, Section S7. The x-axis shows the percent change in the respective input quantity from its baseline. Only parameters with a cost swing of at least $0.05/lge are shown. See the Supporting Information, Section S5.6 for full sensitivity results.
Figure 3. Sensitivity of the next-decade high-temperature solar electrofuel production scenario, individually varying parameters to their today and by 2050 values (with a few exceptions described in the Supporting Information, Section S7). The x-axis is the resulting percent change in each input value. Lines with steeper slopes or larger vertical lengths for a given parameter imply greater sensitivity. WACC is the weighted average cost of capital used to convert capital cost into levelized cost. The high value of DAC electricity demand is 809% of the baseline value, beyond the x-axis limit shown. Parameters with sensitivity less than $0.05/lge are not shown. Note that no parameters associated with energy storage reach this level. See the Supporting Information, Section S5.6 for full sensitivity results.
Electrofuel cost is by far the most sensitive to the capital cost of the DAC, electrolyzer, and renewable electricity generation, introducing cost swings of $0.68/lge, $0.50/lge, and $0.45/lge, respectively. The WACC and the system lifetime, which help determine the annualized capital cost, yield cost swings of $0.45/lge and $0.15/lge, respectively. Fuel synthesis capital cost introduces a modest cost swing of $0.16/lge. Note that these capital cost sensitivities also account for increases in fixed O & M costs, which are modeled as an annual percentage of capital cost.
The system is sensitive to fuel synthesis and electrolyzer efficiency, represented by a steep slope on the plot. However, in both cases, there is limited room for efficiency improvements, leading to modest cost swings of $0.15-$0.16/lge in both cases.
Although no storage-related parameter has a cost swing above $0.05/lge, the CO2 price (set by default to the cost of offsite DACS) and natural gas price introduce costs swings of $0.14/lge and $0.06/lge, respectively.