This study first characterizes greenhouse gas emissions and costs of commercial technologies for blue hydrogen production and then develops technological learning and diffusion models to assess the future costs and evolutionary trajectories of blue hydrogen production without and with tax incentives toward the U.S. Hydrogen Energy Earthshot. A series of parametric analyses are further performed to reveal the dependence of the overall hydrogen production cost on key factors, such as fuel price, carbon capture cost uncertainties, learning rates, and inflation rate.
Current blue hydrogen production
This study adopts state-of-the-art reforming and gasification technologies as a point of reference to explore the evolutionary trends of blue hydrogen production driven by learning-by-doing. The current performance and cost estimates of these technologies are obtained from the recent NETL study7. The majority of hydrogen produced in the U.S. is made via steam methane reforming (SMR). In addition, the cost of blue hydrogen produced by SMR with carbon capture and storage (CCS) is similar to that by autothermal reforming with CCS, but the on-site and life-cycle emissions from the SMR process are less7. This study, therefore, focuses on SMR with CCS for gas-based blue hydrogen production. In the meantime, an oxygen-blown, entrained-flow Shell-type gasifier is employed with CCS for coal-based blue hydrogen production7. Supplementary Table 1 in Supplementary Note 1 summarizes the major techno-economic parameters and assumptions made for blue hydrogen production plants using natural gas and coal resources as the feedstocks, which include the project book lifetime, capacity factor, hourly production capacity, fuel price, and fixed charge rate. In addition, the land and water footprints per unit of hydrogen produced by these plants and the amounts of CO2 sequestration are also reported in Supplementary Table 2. Blue hydrogen plants produce high-purity hydrogen (99.9 vol.%) at the pressure of 6.48 MPa and transport the captured CO2 at the pressure of 15.3 MPa for storage in saline reservoirs, which are typical design conditions. This study reports the cost results in 2018 U.S. real dollars unless otherwise noted.
For the given assumptions, the total levelized cost of hydrogen (LCOH) is $1.64/kg H2 for SMR with CCS and $3.09/kg H2 for coal gasification with CCS. In comparison, the plant LCOH is 88.4% higher for gasification production than the reforming production, which indicates that the integration of SMR with CCS is much more competitive for blue hydrogen. In addition, the on-site stack CO2 emissions from hydrogen production by SMR with CCS are 0.4 kg CO2/kg H2, which is much less than that (1.4 kg CO2/kg H2) from gasification with CCS7. In addition to the stack CO2 emissions, there may be fugitive GHG emissions from various sources at an SMR production plant, mainly from the piping equipment and fittings26. However, fugitive GHG emissions are about 0.05% of the stack GHG emissions26, which indicates that plant methane leakage is not a serious issue.
A hydrogen production plant is decomposed into major subsystems. The components included in each of the subsystems defined for gas- and coal-based production plants are reported in Supplementary Tables 3 and 4, respectively. Figure 1a, b show the distribution by subsystem in the plant LCOH for the two production plants, respectively. The contributions of individual subsystems to the overall production cost are different. Given the gas price of $4.2/GJ, SMR and associated fuel consumption collectively account for 65.9% of the plant LCOH for gas-based production. Please note that at the gas-based hydrogen plant with CCS, the fuel combustion unit generates thermal energy for not only SMR but also the carbon capture process’s solvent regeneration. Given the coal price of $57.3/metric ton, gasification and associated fuel consumption collectively account for 52.1% of the coal-based production. Thus, the progress of individual subsystems in learning will have a different effect on the overall production cost in the future. In addition, the fuel costs account for 50.0% and 14.2% of the plant LCOH for the gas- and coal-based production cases, respectively. Natural gas price is a key factor influencing gas-based blue hydrogen production.
a Distribution of initial levelized cost for gas-based H2 production. b Distribution of initial levelized cost for coal-based H2 production. c Learning curves for coal-based and gas-based H2 production capital and operating and maintenance (O&M) costs without tax incentives. d Learning curves for overall levelized cost of coal-based and gas-based H2 production without and with tax incentives. e Future cost reductions for coal-based and gas-based H2 production with tax incentives.
Currently, hydrogen is mainly produced by SMR without CCS in the U.S., which is often called gray hydrogen. Compared to it, the blue hydrogen production by SMR with CCS can decrease the stack CO2 emission intensity by 96% but increase the LCOH by 55%7. The resulting CO2 avoidance cost by blue hydrogen is $65 per metric ton of CO2. In contrast, the green hydrogen production by polymer electrolyte membrane electrolyzers almost has no stack CO2 emissions but a high LCOH value ranging from $3.0 − 7.5/kg H227. The resulting CO2 avoidance cost by green hydrogen relative to gray hydrogen varies from $212 − 689 per metric ton of CO2, which is much higher than that by blue hydrogen. Obviously, there are tradeoffs in CO2 avoidance cost and emission savings between the blue and green hydrogen production pathways. The details of emission and cost data and CO2 avoidance cost estimation are available in Supplementary Note 2. Please note that the choice of a reference case affects the CO2 avoidance cost.
Future costs of blue hydrogen production without and with tax credits
A blue hydrogen production plant consists of numerous subsystems. However, the maturity status of individual subsystems and their initial installed capacity are different. As a result, learning rates and initial installed capacity vary by subsystem. Thus, a component-based learning curve model is employed to construct a plant-level learning curve based on individual subsystems’ learning rates and initial installed capacity. In addition, the technological learning is evaluated in terms of the cumulative installed capacity of blue hydrogen instead of the number of new hydrogen production plants. To characterize the evolving costs of blue hydrogen produced from natural gas and coal resources in the future, this study first constructs learning curves for the total as-spent capital (TASC) and total operating and maintenance (TOM) cost of individual subsystems at each plant and then establishes the learning curve of the plant LCOH as a function of cumulative production capacity. To construct a learning curve for either the TASC or the TOM, initial installed capacity, initial cost, and learning rate have to first be determined. As discussed above, the initial TASC and TOM of individual subsystems are derived from the NETL study7 and summarized in Supplementary Tables 6–10 in Supplementary Note 3 and Supplementary Tables 11 and 12 in Supplementary Note 4. The initial installed capacity (Supplementary Table 13) and learning rates of individual subsystems are collected mainly from numerous well-established studies and summarized in Table 2, in which bracketed values indicate uncertain ranges related to the base values. Both the initial installed capacity and learning rates vary significantly by subsystem. There are also high uncertainties in learning rates for both capital and O&M costs.
Blue hydrogen production without tax credits. At a global scale, the initial installed capacity of hydrogen production in 2021 was estimated to be 0.31 MMTA for gas-based blue hydrogen and 0.15 MMTA for coal-based blue hydrogen12,28. Figure 1c shows the learning curves for levelized capital and O&M costs and plant LCOH for fossil-based hydrogen production. A comparison between different cost categories over a range of cumulative production capacity indicates that the overall levelized cost of blue hydrogen will still be affected largely by TOM in the future, especially for gas-based production. In addition, a comparison between the two methods implies that SMR with CCS would continue to be more economically competitive for blue hydrogen production than gasification with CCS.
The annual demand for clean hydrogen produced from renewable and decarbonized fossil resources in the U.S. may reach 10 million metric tons of hydrogen per year by 20301. As shown in Fig. 1c, the costs decline via incremental improvements to current technologies when the cumulative production capacity increases. When it reaches 10 MMTA, the capital and O&M costs decrease by 20.0% and 8.3% from the current levels for gas-based production, respectively. There are similar cost reductions for coal-based production. As a result, the plant LCOH decreases to $1.46/kg H2 by 10.7% for the gas-based production and $2.75/kg H2 by 10.9% for the coal-based production after 10 MMTA of hydrogen production capacity. Although experience learned from large-scale deployed projects will help to lower the future costs of blue hydrogen production, it is hard for both reforming and gasification technologies with CCS to reach the cost target of $1/kg H2.
The plant LCOH trends are affected largely by key subsystems’ capital and O&M learning rates, along with feedstock prices. For gas-based production, SMR and CCS are the key subsystems that dominate the plant LCOH, as discussed above. However, as shown in Table 2, there are no reductions gained in O&M costs from deploying SMR (including associated components), pressure swing adsorption (PSA) for hydrogen purification, and CO2 compression. For the given natural gas price of $4.2/GJ, therefore, it is difficult to reach the cost target of $1/kg H2, even when the cumulative production capacity goes beyond 10 MMTA.
Blue hydrogen production with tax credits. Both the 45Q and 45 V tax credits are to promote investment in clean hydrogen technologies and then lower the cost of hydrogen production. For blue hydrogen projects with CCS, however, a taxpayer cannot simultaneously claim both 45 V and 45Q tax credits during a given period. To claim either the 45Q tax credit or the 45 V tax credit, facilities must be placed in service before January 1st, 203329. The credit is available for such qualified facilities for a period. The period of credit availability is common to eligible facilities, regardless of their start-of-service time.
In this study, it is assumed that the captured CO2 is stored in saline reservoirs, which earns a carbon-sequestration credit of $85 per metric ton of CO2 for 12 years. As mentioned earlier, the 45 V tax credit depends on the life cycle emissions of hydrogen production, which include greenhouse gas emissions from plant stacks, fuel supply, electric power supply, and CO2 sequestration or management. The life cycle emissions were estimated by the NETL to range from 3.1 to 8.9 kg CO2-eq/kg H2 for the gas-based blue hydrogen in the 90% confidence interval between the 5th and 95th percentile values and from 3.4 to 8.9 kg CO2-eq/kg H2 for the coal-based blue hydrogen, which is driven mainly by the uncertainty in fuel supply7. The largest contributor among the multiple stages to the life cycle emissions is the fuel supply7. The median estimate of life cycle emissions is 4.6 kg CO2-eq/kg H2 for the gas-based blue hydrogen and 4.1 kg CO2-eq/kg H2 for the coal-based blue hydrogen7, which is close to the threshold value of 4.0 kg CO2-eq/kg H2 required to claim the minimum tax credit for clean hydrogen. Blue hydrogen projects have a fair possibility of earning a 45 V tax credit. Thus, the production tax credit for hydrogen projects is assumed to be $0.6 per kilogram of H2 for 10 years. This assumption is optimistic for blue hydrogen in this study. However, there is no 45 V tax credit if the life cycle emissions of specific blue hydrogen projects are more than 4.0 kg CO2-eq/kg H2. See Supplementary Note 5 for additional information about life cycle emissions and tax credits. Figure 1d shows the learning curves for the plant LCOH for the gas- and coal-based hydrogen production with tax incentives.
Tax incentives lower the plant LCOH of hydrogen production. When the cumulative production capacity reaches 10 MMTA, the overall cost of hydrogen produced by SMR with CCS declines to $1.14 and $1.26 per kilogram of hydrogen produced with the 45Q and 45 V tax credits, respectively. They are 22.2% and 13.7% less than the LCOH without tax incentives, respectively. There are similar cost reductions with gasification-based production. These results indicate that the 45Q tax credit provides more economic incentives for blue hydrogen projects than the 45 V tax credit.
Tax incentives decrease the time-related learning experience necessary to reach a cost target. However, Fig. 1d shows that with either a hydrogen production tax credit or a carbon-sequestration tax credit, it is still hard for coal gasification with CCS to produce blue hydrogen at a cost of $1/kg H2. In contrast, with the carbon-sequestration tax credit claimed for hydrogen projects, the cost of blue hydrogen produced by SMR with CCS approximates the Hydrogen Energy Earthshot, as shown in Fig. 1d.
Learning-by-doing will reduce the cost of hydrogen production for coal- and gas-based blue hydrogen. Figure 1e shows the cost reduction by subsystem and by the 45Q tax credit when the cumulative installed capacity of blue hydrogen reaches 10 MMTA. For blue hydrogen produced from both coal and gas resources, the overall cost reduction will be driven largely by the carbon-sequestration tax credit and the improvement in carbon capture. In contrast, other subsystems, such as SMR and PSA, will make limited contributions because they are mature technologies and have no or limited reductions from an additional 10 MMTA deployment in their future costs. These results indicate the importance of continued support from both public and private sectors for CCS-related research, development, and demonstration programs at federal and state levels.
Time-based diffusion of blue hydrogen production
It is helpful for hydrogen energy planning to explore if certain production capacity and cost targets can be achieved by 2030. A new study reports the cumulative installed capacity of low-carbon hydrogen production over time-based on globally announced, planned, and committed projects through 203030. A diffusion-of-innovation model was established based on the current and future low-carbon hydrogen capacities through 2030 to explore the time-based diffusion of gas-based blue hydrogen over a long-term planning horizon through 2050.
Figure 2a shows the cumulative installed capacity estimates for global low-carbon hydrogen production over time. The gas-based blue hydrogen capacity accounts for 49% of the total low-carbon hydrogen capacity given in Table 1 and is estimated to be 90% in 2030 in terms of the International Energy Agency’s hydrogen project databases28,31. Given the changing shares over time, Fig. 2a also shows a range of cumulative installed capacity for gas-based blue hydrogen in a particular year. The cumulative installed capacity of the global gas-based blue hydrogen may range from 6 to 12 MMTA in 2030, which implies that it would be hard for the blue hydrogen production by SMR with CCS alone in the U.S. to reach 10 MMTA in 2030.
a Diffusion of cumulative installed capacity. b Time-based learning curves of blue hydrogen production cost.
Figure 1c shows the overall plant LCOH as a function of cumulative installed capacity for gas-based blue hydrogen, whereas Fig. 2a shows the cumulative installed capacity over time. Combining them together, Fig. 2b shows the overall plant LCOH of gas-based blue hydrogen production without tax credit over time. The result shown in Fig. 2b implies that for the fuel price and learning rates given in the base case, it would also be difficult for gas-based blue hydrogen to reach the ambitious cost target of $1/kg H2 by 2030 in normal scenarios without aggressive incentives and game-changing technologies.
Sensitivity of blue hydrogen production cost to key factors
Massive deployment of hydrogen projects will lower future costs for clean hydrogen production. Tax incentives for clean hydrogen will further decrease production costs and accelerate the technological evolution toward the Hydrogen Energy Earthshot. However, it is hard for coal gasification with CCS to reach the cost target of $1/kg H2 for clean hydrogen. In contrast, the tax-incentivized production for blue hydrogen by SMR with CCS has the potential to reach the Hydrogen Energy Earthshot. Blue hydrogen projects announced in the U.S. will mainly employ gas-based reformation technologies with CCS31. The future production costs and their evolutionary trends are affected by natural gas price, carbon removal system cost, and learning rates in capital and O&M costs, as well as inflation when the cost is estimated in nominal dollars. In the U.S., natural gas prices are highly volatile. There are also high uncertainties in learning rates for many subsystems, which are shown in Table 2. The sensitivity analysis, therefore, is performed for the gas-based blue hydrogen with a focus on natural gas price, carbon removal system cost uncertainties, learning rates, and inflation rate. In each parametric analysis, other parameters were kept at the base case values given in Table 2 and Supplementary Tables 1, 15, and 16 unless otherwise noted.
Effect of natural gas price. For blue hydrogen produced by SMR with CCS, the feedstock cost accounts for 50.0% of the plant LCOH, with an assumed natural gas price of $4.2/GJ. In the past years from 2017 to 2022, the annual average Henry Hub gas prices ranged from $1.9/GJ to $6.1/GJ32. Thus, it is necessary to examine the economic benefit of low gas prices for blue hydrogen production. Figure 3 shows the effect of natural gas prices on the plant LCOH for hydrogen production without and with tax incentives. Obviously, the plant LCOH is highly sensitive to the natural gas price. Cheap natural gas resources help SMR-CCS decrease the cumulative production capacity necessary to reach the Hydrogen Energy Earthshot.
Effect of natural gas price on future levelized cost of gas-based blue hydrogen production in three cases including base production case, production with a 45V tax credit, and production with a 45Q tax credit.
When the 45Q tax credit is claimed for hydrogen projects, the cumulative production capacity necessary to reach the cost target of $1/kg H2 is 4.9 and 0.6 MMTA if the gas price declines to $3.3/GJ and $2.8/GJ, respectively. As shown in Fig. 3, the initial plant LCOH is already less than $1/kg H2 if the natural gas price is $2.4/GJ. When the 45 V tax credit is claimed, the cumulative production capacity should reach 9.8 and 1.2 MMTA to achieve the cost target when the gas price declines to $2.8/GJ and $2.4/GJ, respectively. Without tax incentives like 45Q and 45 V and increased learning rates, however, it is still difficult for this production method, even with cheap natural gas resources to reach the Hydrogen Energy Earthshot.
It is also worth noting that the gas-based hydrogen industry may have a sizable effect on the natural gas markets in the U.S., depending on the scale of blue hydrogen production in the future. For example, the production of 10 MMTA hydrogen by SMR with CCS would consume 1.9 billion GJ of natural gas per year, which is equivalent to about 17% of the national industrial natural gas consumption in 202233.
Effect of carbon removal system cost uncertainties. There are uncertainties in the process and project contingencies of two CO2 removal systems employed for producing low-carbon hydrogen from natural gas resources. Such uncertainties affect the TASC and LCOH of a hydrogen production plant. The process contingency depends on the maturity level of a technology, whereas the project contingency depends on the availability of site-specific project details. In the base case, the process contingency is 18% of the bare erected cost (BEC) for the Cansolv system and 0% for the MDEA system, while the project contingency is 25% of the sum of BEC, engineering, construction management, home office and fees, and process contingency and 25% for the Cansolv unit and 20% for the MDEA unit7. A parametric analysis is then conducted to reveal the collective impacts of uncertain processes and project contingencies, which take into account low and high contingencies. In the low contingencies scenario, the process contingency is 10% for the Cansolv system and 0% for the MDEA system, while the project contingency is 10% for both the CO2 removal systems. In the high contingencies scenario, the process contingency is 40% for both the CO2 removal systems, while the project contingency is 30% for both the CO2 removal systems34.
As shown in Fig. 4, the uncertainties in carbon removal system cost estimates have a sizable effect on the hydrogen production plant’s TASC. As a result, the plant LCOH varies from $1.45 to 1.48/kg H2 at the cumulative installed capacity of 10 MMTA. To reach the cost of $1.46/kg H2, the cumulative installed capacity requirements vary from 7 to 16 MMTA. These results imply that cost uncertainties in carbon removal systems may result in pronounced variations in the estimation of the cumulative installed capacity necessary to reach a cost target.
The light blue shading area represents the uncertainties in the levelized cost of hydrogen production, which are driven by the uncertain cost estimates of carbon removal systems. The orange dash lines represent that the levelized cost of hydrogen varies from $1.45 to 1.48/kg H2 at the cumulative installed capacity of 10 MMTA, while the blue dash lines represent that to reach the cost of $1.46/kg H2, the cumulative installed capacity requirements vary from 7 to 16 MMTA.
Effect of learning rates. Learning rates directly drive future cost trends. In particular, the O&M learning rates of SMR, PSA, and CO2 compression largely influence the pace of cost reductions toward the Hydrogen Energy Earthshot as their costs and associated fuel or electricity consumption collectively dominate the plant LCOH. In the base case, the O&M learning rates are zero for the three subsystems. However, Table 2 shows that there are uncertainties in learning rates, which can vary by 50% or more relative to the base values for some subsystems. Given such high uncertainties, it is important to examine the sensitivity of future cost trends to learning rates.
Additional scenarios are explored to examine the effect on the overall LCOH of increases in both the capital and O&M learning rates of individual subsystems with an emphasis on the increased O&M learning rates for SMR, PSA, and CO2 compression. In these scenarios, the capital and O&M learning rates are elevated for individual subsystems to be 25% and 50% higher than the base values, except for SMR, PSA, and CO2 compression. The O&M learning rates are increased for the three subsystems to 5% and 10% on an absolute basis. Figure 5 shows the sensitivity of the plant LCOH to the increased learning rates.
a LCOH under two boundary scenarios of learning rates. b LCOH under the range of 100%–150% time-based learning rates, except for O&M cost learning rates, which are equal to 5%–10% for SMR, PSA, and CO2 compression. Note to Fig. 5: P1 means a percentage relative to the base learning rate, whereas P2 means the learning rate on an absolute basis. Note to abbreviations: CCS means carbon capture and storage; LCOH means levelized cost of hydrogen; LR means learning rate; O&M means operating and maintenance; PSA means pressure swing adsorption; and SMR means steam methane reforming.
The increases in learning rates, especially the O&M learning rates of SMR, PSA, and CO2 compression, obviously lower the cumulative production capacity necessary to reach a cost target. The results shown in Fig. 5, however, also imply that without tax incentives for clean hydrogen, it would still be challenging for blue hydrogen produced from expensive natural gas resources to reach the Hydrogen Energy Earthshot by 2030, even if the progress in learning is to accelerate substantially. If the O&M learning rates of SMR, PSA, and CO2 compression reach more than 5%, massive deployment of blue hydrogen projects claimed with 45Q tax incentives can decrease the plant LCOH to $1/kg H2. Figure 5a,b show that with an O&M learning rate of 10% for the three subsystems, the breakeven cumulative production capacity is 20 MMTA or more, which is also affected by other subsystems’ learning rates.
Effect of inflation rate. In general, this study estimates the cost of hydrogen production in real dollars. When the cost is estimated in nominal dollars, however, both the initial and future LCOH estimates vary with the inflation rate as it affects the discount rate, fixed charge rate, and levelization factor. A parametric analysis was further performed for the inflation rate to quantify its effect on the evolving cost of gas-based blue hydrogen production toward the Hydrogen Energy Earthshot. Figure 6 shows the learning curves of blue hydrogen production with inflation. Figure 6a, b show that at a given level of cumulative installed capacity, the LCOH in nominal dollars increases when the inflation rate increases from 1% to 3%. As a result, blue hydrogen production may not reach the cost target of $1/kg H2 for both scenarios without and with a 45Q tax credit even when the cumulative installed capacity reaches 30 MMTA. Figure 6c further shows that with an inflation rate of 3%, the future LCOH may get close to the cost target when cheap natural gas resources are used as the feedstock to produce blue hydrogen with a cumulative installed capacity of up to 30 MMTA.
a Levelized cost of hydrogen production with a gas price of $4.2/GJ and without a 45Q tax credit. b Levelized cost of hydrogen production with a gas price of $4.2/GJ and 45Q tax credit. c Levelized cost of hydrogen production with a 3% inflation rate and 45Q tax credit.
Figure 6a,b also compare the learning curves of blue hydrogen production between the two scenarios without and with inflation. As shown in Fig. 6a for the scenario without a 45Q tax credit, the reduction in hydrogen production cost from deploying the cumulative installed capacity of 10 MMTA can be offset by an inflation rate of 1%. There is a similar result at the cumulative installed capacity of 5 MMTA for the scenario with a 45Q tax credit, as shown in Fig. 6b. All these results imply that inflation would remarkably raise challenges for blue hydrogen production to reach the Hydrogen Energy Earthshot in the near future.
