There’s a Lot to Celebrate in Treasury’s Clean Hydrogen Tax Credit Proposal
This morning, the Department of the Treasury released a notice of proposed rulemaking (NOPR) on the clean hydrogen production tax credit, and it incorporates many key recommendations from climate policy advocates and a segment of the hydrogen industry. Treasury’s approach could help avoid a scenario where subsidies support emissions-intensive hydrogen that undercuts climate progress.
If you weren’t tracking this issue, the first thing to know is that the Biden Administration’s landmark climate law (the Inflation Reduction Act, or IRA) bet big on hydrogen as a climate solution with a lucrative tax credit for producing hydrogen. Treasury’s NOPR proposes details about how the hotly debated credit (Section 45V of the tax code) will work.
When hydrogen is burned (or used in a fuel cell), it releases energy but no CO2. So, there’s an opportunity to replace some fossil fuels with hydrogen, particularly for energy needs that would be hard to meet using only renewables plus batteries (e.g., balancing the electric grid during prolonged periods of low renewable output). But creating hydrogen can release significant greenhouse gas emissions (GHGs). The IRA’s smallest subsidy goes to hydrogen with an emissions intensity of 2.5 to 4 CO2e/kg H2. And its most generous subsidy is reserved for hydrogen with an emissions intensity of less than 0.45 CO2e/kg H2.
The most common way to make hydrogen — steam methane reforming — uses methane as a feedstock and releases CO2 as a byproduct. This process results in 10–12 kg CO2e/kg H2. Carbon-capture technology can reduce these emissions (4.6 kg CO2e/kg H2 with a very high carbon-capture rate of 96.2%). But, from a climate perspective, the worst way to make hydrogen is actually fossil-fuel-powered electrolysis: burning fossil fuels to make electricity, and then using the electricity to split water into hydrogen and oxygen. Burning natural gas to make electricity for hydrogen production emits roughly 20 kg CO2e/kg H2. In contrast, the most climate-friendly way to make hydrogen is to power your electrolyzer with zero-emissions electricity from a renewable resource. This results in hydrogen with an emissions intensity of 0 kg CO2e/kg H2.
By establishing commonsense measurement rules for the emissions intensity of electrolytic hydrogen, Treasury’s NOPR goes a long way towards ensuring that we don’t accidentally subsidize fossil-fuel-powered electrolysis. Without these rules, it would be easy to confuse fossil-fuel-powered electrolysis with renewables-powered electrolysis. If that were to happen, the most significant climate law in U.S. history would have the perverse effect of subsidizing hydrogen that causes twice as many emissions as the old method of steam methane reforming.
In many situations, it isn’t hard to measure the emissions intensity of hydrogen. But this emissions accounting is especially tricky for electrolyzers that draw power from the regional electric grid. When an electrolyzer plugs into the grid and begins to draw power, how can we know whether the power is coming from a wind farm or a coal plant? For this critical question, the NOPR adopts the preferred solution of environmentalists and a segment of the hydrogen industry: the so-called “three pillars” approach.
The three pillars are additionality, hourly matching, and deliverability. By applying these three principles — all of which describe restrictions on how electrolyzers can use their electricity purchases to qualify for the tax credit — it becomes much easier to identify bogus claims of low-emissions hydrogen. Here’s why:
Additionality (or, as the NOPR calls it, Incrementality)
Treasury proposes that an electrolyzer can use electricity purchases to qualify for the IRA’s tax credits only when the generator began commercial operations no more than three years before the electrolyzer was placed into service. Added capacity at existing plants within the last three years also qualifies. This restriction aims to prevent the problem of “resource shuffling.” Imagine a new electrolyzer comes online and contracts for electricity with a renewable resource that, until then, was producing electricity for a different customer. If the other customer is agnostic about the emissions intensity of its electricity, it may turn to a fossil-fuel-fired generator. From a system-wide perspective, the electrolyzer would have reshuffled the renewable generation on paper while adding significant fossil-fuel-fired generation and no new renewable generation.
Under these circumstances, it would make little sense to allow the electrolyzer to point to the contracted renewable generation to prove that it deserves tax credits based on the emissions intensity of its hydrogen. The electrolyzer set in motion a series of events that resulted in the same emissions as if it had contracted with the fossil-fuel-fired generator for all of its electricity.
Treasury’s three-year rule will help solve this problem by ensuring that electrolyzers cannot divert low-GHG electricity from long-existing generators. This same restriction was already adopted by the European Union.
Hourly Matching
The NOPR would place timing restrictions on electricity purchases that electrolyzers use to qualify for the tax credit. Until the end of 2027, electrolyzers can establish the emissions intensity of their hydrogen by contracting for electricity generated within the same year as the hydrogen was produced. Starting in 2028, the electricity will need to have been generated within the same hour as the hydrogen. This eventual shift to hourly matching will help ensure that we don’t subsidize fossil-fuel-powered electrolysis, but it is necessary to understand some basics about the electric grid to see how.
Given the realities of grid operation, the most accurate way to measure emissions from grid-connected electrolyzers involves looking at the emissions intensity of the “marginal” generator serving the local grid at the moment of hydrogen production. The marginal generator is whichever generator the grid operator would ask to increase its output to meet additional demand for electricity. The identity of the marginal generator — and thus the marginal generator’s emissions intensity — changes frequently throughout the day. The comments we submitted with WattTime to Treasury last December included Figure 1. This figure depicts the marginal emissions rate in the Southwest Power Pool (a regional power grid in the central United States) over one month, as modeled by WattTime. It reveals that the marginal emissions rate can swing back and forth from zero lbs CO2/MWh to over 1,400 lbs CO2/MWh repeatedly throughout a single day.
This background leads to two critical points. First, the emissions from the marginal resource would be avoidable if the electrolyzer didn’t run; therefore, the electrolyzer causes the emissions of this marginal generator. Second, when a renewable resource comes online and allows the marginal resource to ramp down, the renewable resource is responsible for avoiding the emissions that the marginal generator would have emitted.
And so, because the marginal resource can change so quickly, emissions accounting becomes inaccurate when there’s a large time gap between the electricity that an electrolyzer consumes and the electricity that it purchases. If an electrolyzer draws power from the grid at a time when the marginal emissions rate is higher than the marginal emissions rate when the contracted-with generator injects power, the electrolyzer induces more emissions than the generator avoids. But this problem is less likely to occur when the matching happens over a tighter interval (like an hour) as compared to a longer interval (like a year), because the marginal generator is less likely to change when little time has passed. In contrast, if an electrolyzer merely needs to buy low-GHG electricity that was generated within the same year as the electrolyzer’s power consumption, there is a real risk that the marginal emissions rates would diverge.
Although Treasury proposes to allow annual matching until 2028 — at which point hourly matching would be required — an analysis from RMI concluded that a phased-in approach would be good policy. The analysis found that transitioning from monthly matching to hourly matching in 2028 would accelerate the development of electrolytic hydrogen while causing relatively few additional emissions as compared to immediate hourly matching. Treasury’s proposed approach is quite similar to the EU’s policy of monthly matching until 2030 followed by hourly matching.
Deliverability
Treasury proposes to restrict electricity purchases to the same regional grid to ensure “deliverability.” Deliverability refers to whether the contracted-for electricity can travel from the generator to the electrolyzer, given their locations within the transmission network. As with hourly matching, this limitation is necessitated by differences in marginal emissions rates.
Without deliverability, an electrolyzer might consume power in a region where the marginal resource is a fossil-fuel-fired plant (thus inducing the emissions of that plant) while contracting with a generator located somewhere where renewables are on the margin. The result would be fossil-fuel-powered electrolysis in the first region, while the renewable generation would not avoid any emissions in the second region (because renewables were on the margin there).
Figure 2, created by WattTime based on their modeling, is a snapshot of the spatial variation in emissions rates of marginal resources at a moment in time. It shows that the marginal emissions rate can diverge sharply even between two regions that are geographically proximate.
Although transmission constraints can exist within a single regional grid, Treasury’s proposal to require same-grid transactions is a good first approximation of deliverability. The EU has adopted an analogous deliverability requirement.
Treasury is now accepting comments on the NOPR. If Treasury finalizes a rule with the proposed approach, it would lay the groundwork for a growing clean hydrogen industry that can help address a range of decarbonization challenges.
