As hydrogen hype continues to mount, two papers were published recently on the climate impacts of “blue” hydrogen — that is, hydrogen from natural gas with the addition of carbon capture and sequestration. I assessed the second, by Bauer et al, in a pair of articles, mostly looking at their underlying assumptions and conclusions, both of which I found to be challenging. As I wrote, it appeared to have been in the works for a while and rushed to pre-publication, possibly with additional signatories who felt compelled to climb on, in response to the first paper, one by Mark Z. Jacobson and Robert Howarth.
Their paper, “How green is blue hydrogen?,” was peer reviewed and published in Wiley’s open access journal Energy Science & Engineering, which has an impact factor of 4.07. As a note, it is a pay-to-publish journal, as many are these days, but is not a predatory journal with low standards. The Bauer et al paper was in ChemRxiv, a pre-publication journal without peer review allowing early access to papers, also a common model in academia at present, and not necessarily indicative of quality. Given the peer review, however, Jacobson and Howarth’s paper should be considered to be slightly more reliable by this metric.
In assessing the Bauer et al paper, I did not question their assumptions on their life-cycle assessment process, something that they tout as being a key part of the expertise and differentiation that they bring. They assert directly that the Jacobson and Howarth paper was inferior in this regard.
“Finally, the more recent analysis does not follow best practices in LCA as it, for example, takes into account neither GHG emissions associated with capital goods nor those originating from transportation and geological storage of CO2; and it relies on data for natural gas supply only in the US context.”
It’s worth noting that the first two points in this would increase the greenhouse gas emissions, and are a fraction of the GHG emissions related to upstream methane emissions in both models, so they are immaterial to the larger point. It’s good that Bauer et al include them, but immaterial as a complaint against Jacobson and Howarth.
The last complaint, that it is US-specific, may or may not be accurate, as one of the references is to a study of 16 gas-producing regions being published in a book I don’t have access to (and neither do the authors of the Bauer et al paper, as far as I know). Howarth has testified regarding atmospheric methane and extraction globally. Certainly, Howarth’s publication history is US-centric, so this may be a fair comment, but as we’ll see, Bauer et al leverage this to arrive at radically different numbers, and for a great deal of their publication focus on US results as well.
The approach I’ll take in this piece is the same as in my initial assessment of the Bauer et al paper, in that I’ll extract key portions, add context where necessary, and discuss them. Further, where there is significant difference between the papers, I’ll highlight it and examine which arguments I think hold most merit.
“We undertake the first effort in a peer-reviewed paper to examine the lifecycle greenhouse gas emissions of blue hydrogen accounting for emissions of both carbon dioxide and unburned fugitive methane.”
This strikes me as relatively humble. They acknowledge that this hadn’t been done before, saw the need, and took it upon themselves, as experts in climate change and solutions, to do the first assessment and publish it for discussion. They weren’t doing a hardcore LCA process per Bauer et al and not claiming that, but were looking at the material elements.
“Our analysis assumes that captured carbon dioxide can be stored indefinitely, an optimistic and unproven assumption.”
As always when reviewing Jacobson’s work, I find conservative estimates and choices. In the major study of 143 countries and how they could deliver all energy services with renewables by 2050, there are no unknown or in-development technologies chosen. They note later in the paper that the majority of carbon “sequestration” is for enhanced oil recovery, and that not only can CO2 escape from these more porous facilities, it does, so they are simply not calculating this, but assuming permanent storage.
“In this analysis, we consider emissions of only carbon dioxide and methane, and not of other greenhouse gases such as nitrous oxide that are likely to be much smaller.”
This is another conservative choice, as nitrous oxide (N2O) has a global warming potential (GWP) of 265 compared to methane’s 86. And when hydrogen is burned, nitrous oxide is created as well. Burning things in our atmosphere, made up of 78% nitrogen and 21% oxygen, causes a chemical reaction in which some of the nitrogen and oxygen combine into nitrous oxide. The last point to make about nitrous oxide is that it’s also bad for the ozone layer. In general, it’s a bad idea to be creating it if we don’t have to, which is another reason to avoid burning things unless we need to. Most of our burning of fuels for energy needs to stop. However, the major anthropogenic sources of nitrous oxides are agriculture, with combustion playing a smaller role.
It’s an obvious place where a follow-on study would reasonably include nitrous oxide emissions due to their high GWP. And full lifecycle consideration of hydrogen as a combustion fuel would necessarily need to consider them as well.
It’s worth noting that the purportedly superior LCA approach of Bauer et al is entirely silent on nitrous oxide, and so should be considered as differently flawed from an LCA perspective, something that they might want to consider.
Subset of Table 1 from Jacobson and Howarth’s paper
It’s also worth noting that the capture rate of CO2 being considered is at the top end of the majority of implementations that the more recent paper by Bauer et al use. That paper had identified a range of 50% to 85% from the vast majority of existing implementations, and then noted that over 90% was being achieved in a few locations, albeit with significant unstated caveats.
This is, once again, a conservative choice based on what is being done in the real world in the majority of cases, and does not unduly burden the analysis with the bottom end of 50%, although many attempted power station CO2 capture systems often performed very poorly.
This is part of the set of calculations performed to assess the three scenarios. The first was for no CO2 capture at all. The second was CO2 capture only from that created from the SMR process. This is similar to capturing the CO2 that bakes off of limestone as it is converted to quicklime for cement, but not capturing the emissions from burning natural gas or coal to create the necessary heat. The third is for also capturing the CO2 emissions from the natural gas generation unit powering the process.
One of the dirty secrets of carbon capture is that it’s an energy intensive process. Whether it’s Global Thermostat’s sorbents from Corning in a batch process or Carbon Engineering’s continuous process, you need to put a lot of heat energy into the capture medium to get the CO2 back out. And while Global Thermostat intentionally tries to use waste industrial heat, the rest of the CO2 capture world just burns natural gas. That’s certainly what Carbon Engineering does, and while it has one carbon capture technology for its air carbon capture, it has to use two completely different technologies in succession to capture the half ton of CO2 it creates burning natural gas for every ton of CO2 it gets from the atmosphere. As I wrote years ago, carbon capture is expensive because physics.
It’s worth pointing out that the 92% figure that Bauer et al cite for Petra Nova excludes the gas cogen unit built specifically to power the process, and the emissions of that cogen unit were not counted in the 92%, and in fact only represented the periods when the capture facility was actually operating. It’s also worth noting that it was scaled up from its original target of a 60 MW coal facility to a 250 MW coal facility because “the original design of a 60 MWe facility was deemed insufficient to meet the CO2 needs of the oilfield” for enhanced oil recovery.
As a reminder, virtually all carbon ‘sequestration’ done today is for enhanced oil recovery, and for every ton of CO2 injected into tapped out oil wells, 0.25 to one ton of crude is recovered. When used as intended, that crude produces more CO2 than was injected, up to three times as much. Petra Nova and Boundary Dam in Saskatchewan, both cited by Bauer et al, were both enhanced oil recovery CO2 providers and both failed economically.
All of this is to say that I find Jacobson and Howarth’s 85% figure to be more compelling than Bauer et al’s, conservative and still actually in favor of “blue” hydrogen.
Bauer et al are not ignorant of this, by the way. They state that current “blue” hydrogen test sites are only capturing 50–60% of plant-wide emissions, yet use 93% CO2 capture at hydrogen steam reformation plants in their modeling, and assert that approaching 100% is likely.
And so ends part one of my assessment of Howarth and Jacobsons “blue” hydrogen LCA, comparing and contrasting it with Bauer et al’s. So far, Jacobson and Howarth have an edge. But in the next piece, some truly remarkable variances between the two studies emerge which make it clear, at least to me, which paper to place credence in.
Appreciate CleanTechnica’s originality? Consider becoming a CleanTechnica Member, Supporter, Technician, or Ambassador — or a patron on Patreon.
PlatoAi. Web3 Reimagined. Data Intelligence Amplified.
Click here to access.
