This group brings together the best thinkers on energy and climate. Join us for smart, insightful posts and conversations about where the energy industry is and where it is going.


How Much Do Ultra-Supercritical Coal Plants Really Reduce Air Pollution?

Zouxian coal power plant in China.


To understand the difference between subcritical, supercritical and ultra-supercritical power generation technology on the air pollutant emissions from a coal-fired power plant, the most important thing to know is this: which type of steam cycle is used has no impact on the emissions per tonne of coal burned.

Taking the example of sulphur dioxide (SO2) emissions, emissions per tonne of coal depend solely on the amount of sulphur contained in the coal, essentially all of which is oxidized into SO2 during combustion, ending up in the raw flue gas.

For example, for typical “low-sulphur” coal containing 0.5 per cent of sulphur when fed into the boiler, every tonne of coal will contain 5 kilograms of sulphur. When burnt, this sulphur turns into 10 kilograms (kg) of SO2. (Every sulphur atom joins with two oxygen atoms to produce one SO2 molecule which is twice as heavy as a sulphur atom.)

The only difference between different steam cycles in terms of emissions is how much power they can generate from one tonne of coal.

A typical new subcritical plant will have a thermal efficiency of 38 per cent, meaning that 38 per cent of the thermal energy contained in the fuel is converted into electrical energy fed into the grid.

A supercritical plant will have an efficiency of maybe 42 per cent and a typical ultra-supercritical plant will achieve around 44 per cent (designs going up to 47 per cent are being developed).

Moral of the story: Emissions regulation matters a lot, whether a plant is ultra-supercritical matters little

This means that a 1000 megawatt (MW) coal-fired plant using subcritical technology will need to burn coal at a thermal input rate of 1000 MW / 38 per cent = 2630 MW-thermal to generate its full output. This corresponds to 410 tonnes of coal per hour, assuming a typical calorific value of 5500kcal/kg, and 4100kg/h of SO2 in raw flue gas.

If the plant uses ultra-supercritical technology, it needs thermal input of 1000 MW / 44 per cent = 2270 MW-thermal. As a result, it burns 350 tonnes of coal per hour, or 14 per cent less than the subcritical plant and generates 14 per cent less SO2.

If the plant is not equipped with SO2 emission control technology, that’s the end of the story.

However, if the environmental regulators require the plant to meet SO2 emission limits that cannot be met without installing SO2 control devices, the plant will have to make additional investments.

Stringent regulation

In essentially all countries except the US, SO2 emission limits are set in terms of SO2 concentrations in flue gas. The project developer will have to design a control device that removes enough of the SO2 from the flue gas to get below the limits.

Some of the toughest limits for SO2 emissions are found in China, where flue gases from coal-fired power plants are not allowed to contain more than 35 milligrams of SO2 for every cubic meter of dry flue gas.

The untreated flue gas from the example plants above will contain about 1200mg/m3 of SO2. Therefore, the plants will have to install SO2 control devices that remove about 97.5 per cent of the SO2 contained in untreated flue gas.

I hope you’re not too shocked that coal advocates are not mainly motivated by health concerns

The difference between subcritical and ultra-supercritical technology is that the total amount of flue gas emitted from the ultra-supercritical plant is about 14 per cent smaller, and hence the capacity of the SO2 control device can be about 14 per cent lower, resulting in savings in investment and operating costs. Resulting SO2 emissions associated with a given emission standard will also be about 14 per cent lower.

The same logic applies to the emissions coal plant (NOx), particulate matter (PM), mercury and other heavy metals. The air quality and health impacts are directly proportional to emissions.

Moral of the story: Emissions regulation matters a lot, whether a plant is ultra-supercritical matters little.

So why are the coal industry and its advocates always going on about ultra-supercritical coal plants and not about emissions regulation?

Simple: ultra-supercritical plants are usually more profitable than subcritical plants, since they have lower fuel and other operating costs.

Stringent emission regulation, in contrast, increases both investment and operating costs. I hope you’re not too shocked that coal advocates are not mainly motivated by health concerns.

Coal and gas plants

It is worth noting that Australia, the main peddler of “High Efficiency Low Emissions” (HELE) coal plants along with Japan, hasn’t even required flue gas desulphurisation equipment on its own coal plants, making them some of the dirtiest in the world.

Below is a simple graph illustrating the effect of emissions regulation versus type of steam cycle on SO2 emissions:

The chart below shows a comparison between coal and gas plants following the same Chinese emission standard.

SO2 and particle emissions from gas are a tiny fraction of those from coal, while NOx emissions are similar. It would be technically easy for the gas plant to go a lot lower but this is what current standards require.

Editor’s Note

Lauri Myllyvirta works for Greenpeace East Asia. This article was first published on and is republished here with permission.

Original Post

Lauri Myllyvirta's picture

Thank Lauri for the Post!

Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.


Darius Bentvels's picture
Darius Bentvels on Jun 29, 2017 7:46 am GMT

The type of the burning process used has a major influence on the amount of CO2 (GHG) per KWh. The metric in the post above:”CO2 per kg coal burned”, is not relevant.

Old coal plants usually have efficiencies of ~33% while the new ultra-super-critical plants have efficiencies of ~44% while also keeping a much higher efficiency at lower power power output range. With old coal plants the already low efficiency goes down must faster at lower outputs.
That implies >25% less CO2 per KWh

Also >25% less SO2, but that is less important in advanced countries as it has all to be filtered out of the flue gas anyway.

The lower burning temperatures in the over-oxidized environment of the circulating fluidized bed burning process in the furnace of ultra-super-critical plants, have other important benefits in addition to the lower (steel, etc) wear (=less maintenance).
It implies that hardly any of the far more health damaging (toxic) NOx’s are created. Important as those NOx’s are difficult to filter and continue to exist long.

Particulate Matter (PM’s) are easy to filter. In advanced countries, old coal plants also (have to) do that.

SO2 and PM’s from gas are only a tiny fraction of those from coal in backwards countries with little regulations.

Bob Meinetz's picture
Bob Meinetz on Jun 29, 2017 2:57 pm GMT

Moral of the story: Emissions regulation matters a lot, whether a plant is ultra-supercritical matters little

A more-accurate moral, Lauri, would be: “Emissions regulation matters little for limiting SO2 emissions; ultra-supercriticality matters little for limiting all emissions; replacing coal with modern nuclear plants would eliminate all emissions, making Greenpeace’s anti-science opposition to it not only ignorant, but dangerous.”

Roger Arnold's picture
Roger Arnold on Jun 29, 2017 6:37 pm GMT

The type of the burning process used has a major influence on the amount of CO2 (GHG) per KWh. The metric in the post above:”CO2 per kg coal burned”, is not relevant.

Not true. In coal-fired plants, only the net thermal efficiency of generation affects the amount of CO2 per KWh. The “type of burning process” may affect other things, but the ratio of CO2 produced to thermal energy of combustion is basic carbon chemistry. The author’s calculations about CO2 per KWh based on thermal efficiency are valid; the higher thermal efficiency possible in supercritical steam plants do reduce CO2 produced per kWh, but not dramatically. The reduction isn’t dramatic, because the improvement in thermal efficiency isn’t dramatic.

I agree with the rest of the statements in your comment.

Darius Bentvels's picture
Darius Bentvels on Jun 29, 2017 8:56 pm GMT

There are a few other important benefits of the low burning temperatures in the over-oxidized environment of the circulating fluidized bed burning process in the furnace of ultra-super-critical plants.

That over-oxidized low temperature environment facilitates the burning of coal mixtures with waste and biomass, while generating far less toxic emissions.
Many of the plants here (NL, Germany) use that feature.

Furthermore, also mainly thanks to the low burning temperatures, those plants can down regulate much faster and deeper without losing so much efficiency.

It makes them suitable in environments with an high share of (variable) renewable, like gas plants. They emit ~30% more CO2 per KWh compared to gas.
Though with burning a mixture of coal with 30% biomass, CO2 emission per KWh is roughly equal.

Darius Bentvels's picture
Darius Bentvels on Jun 29, 2017 9:10 pm GMT

We mainly agree.
The target of such plant is the production of KWh. Hence the important metric is not gram CO2 per kg coal burned, but gram CO2 per KWh produced.

Gram CO2 per KWh produced is also the metric published by environmental agencies such as German UBA.

The author compares with coal plants with an efficiency of 38%. However old coal plants usually have an overall efficiencies of only ~33%.
So the new plants emit ~25% less CO2 per KWh.

I very much doubt whether her lowest diagram is correct regarding NOx emissions.

Darius Bentvels's picture
Darius Bentvels on Jun 30, 2017 3:22 pm GMT

Sorry, but your statement: “modern nuclear plants would eliminate all emissions”
is wrong.
According to the study of Stanford professor Jacobson, nuclear emits 9 – 16 times more CO2 per KWh produced than wind. A result confirmed by this publication in The Conversation.

Common sense tells that nuclear emits ~5 times more CO2 per KWh than wind as nuclear is ~5 times more expensive.

Engineer- Poet's picture
Engineer- Poet on Jun 30, 2017 9:36 pm GMT

We have proof that Jacobson is a liar, a professional anti-nuclear propagandist.  He makes up things like 1300 GW of hydropower capacity coming from nowhere, without which his scheme just doesn’t work.

Bob Meinetz's picture
Bob Meinetz on Jul 1, 2017 3:48 am GMT

Bas, you’re not still quoting Jacobson, are you? Two weeks ago 21 prominent scholars stuck a fork in him. He’s done:

In a long-awaited article published this week in The Proceedings of the National Academy of Sciences — the same journal in which Professor Jacobson’s manifesto appeared — a group of 21 prominent scholars, including physicists and engineers, climate scientists and sociologists, took a fine comb to the Jacobson paper and dismantled its conclusions bit by bit.

“I had largely ignored the papers arguing that doing all with renewables was possible at negative costs because they struck me as obviously incorrect,” said David Victor of the University of California, San Diego, a co-author of the new critique of Professor Jacobson’s work. “But when policy makers started using this paper for scientific support, I thought, ‘this paper is dangerous.’”

If common sense tells you emissions are somehow related to cost, common sense tells me you’re coming unhinged.

Jarmo Mikkonen's picture
Jarmo Mikkonen on Jul 1, 2017 9:14 am GMT

Jacobsen is debunked even in the wikipedia article about him:

In 2012 Heath and Warner from Yale University and the National Renewable Energy Laboratory analyzed all the previous work on the total life-cycle greenhouse-gas emissions of nuclear energy and did not arrive at the same nuclear power values or judgements that Jacobson has. Determining instead that nuclear is “comparable [to] renewable energy” systems, in terms of the total life cycle carbon footprint and that the most supported value for nuclear is 12 g/kWh. While Jacobson’s results are at the higher end of the two extreme poles of peer-reviewed calculations that the IPCC deemed worthy of consideration (1-220 g/CO2eq/kWh), the Intergovernmental Panel on Climate Change(IPCC) regard Warner and Heath’s methodology as the most credible and thus also report that the nuclear power emission is 12 g/kWh, which is comparable to wind energy.

Darius Bentvels's picture
Darius Bentvels on Jul 1, 2017 11:07 pm GMT

Jacobson debunked the attack that you linked. E.g. here.

As he also states; their is no necessity to use so much hydro with all its (NIMBY, etc) hassles.

Using PtG with storage in deep earth cavities is cheaper, more flexible and faster to implement (a.o. little NIMBY).
I estimate that USA has enough deep earth cavities to store a year of energy consumption.
The Germans preferred Power-to-Gas above more hydro.
They will have 2GW pilot capacity in 2022 and start full scale roll-out in 2024. This page gives an overview, their pilots, etc:

Darius Bentvels's picture
Darius Bentvels on Jul 1, 2017 11:25 pm GMT

Jacobson is in good company regarding his 100% renewable study results:

French scientific govt institute ADEME concluded regarding 2050, that 100% renewable would be only 4% more expensive than the cheapest situation, being 80% renewable:

It concerns electricity but near all energy consumption is expected to become electricity.

German think tank Agora concluded similar.

Denmark targets 100% renewable regarding electricity in 2040 and regarding all energy consumption in 2050.
Denmark is now at ~60%, most (~70%) being wind.


Nathan Wilson's picture
Nathan Wilson on Jul 2, 2017 3:21 pm GMT

The difference is that while adding windpower to a fossil fuel dominated grid can reduce fossil fuel consumption, the fossil fuel plants, and most of the fossil fuel consumption are still needed, since the wind does not always blow.

Nuclear power and nuclear plants can be a one-for-one replacement for coal plants and coal use.

(Yes, I know we are often promised that 100% renewable grids are coming, but many decades into the the renewable revolution, there are still none in the world; variable renewables create grid problems which are most easily solved with fossil fuel).

Jesper Antonsson's picture
Jesper Antonsson on Jul 3, 2017 11:18 am GMT

Jacobson’s study has been debunked and is based on very strange assumptions, such as adding burning cities from nuclear wars.

If you check out surveys of LCOE studies, such as OpenEI, nuclear power from a life cycle perspective is as good as wind and better than solar. But nuclear wins out easily if you include grid and backup/storage.

Jesper Antonsson's picture
Jesper Antonsson on Jul 3, 2017 11:21 am GMT

You’re impressively well educated in coal generation and often speak of its advantages and the superior efficiency of new plants. It’s quite a sharp contrast to your tendency to serve us junk science on nuclear power.

Jarmo Mikkonen's picture
Jarmo Mikkonen on Jul 3, 2017 3:01 pm GMT


Deutsche Bank evaluated German Energiewende and crunched some numbers. Here is their scenario for a situation where renewables produce 60% of primary energy, the rest is fossil fuels.

The third scenario comes very close to the long-term target set by the German government: a 50% reduction in primary energy consumption and an increase in the renewables proportion to 60%. As in the other scenarios, the average payment per kilowatt hour is an important lever: if this could be successfully lowered to 8 cents, the costs (just) for the energy from renewables would amount to just under EUR 90 billion p.a. The capital investment costs to reduce energy consumption in this case would probably be very high indeed. Of course, there would be lower ongoing energy costs on the other side of the equation; the overall outcome in economic terms would depend on many factors, such as energy prices, technical progress, etc.

Furthermore, in this scenario, the costs of upgrading vast sections of the heating market (more than 40 million homes in the housing stock) and of the transport sector to electricity would be particularly significant. In a world in which 60% of the total energy supply were based on renewables, there would then also be a further very high level of costs for storage, capacity reserves and load management to be added to the other blocks of costs (networks, other energy sources and infrastructure).

The scenarios are only intended as food for thought. It is useful to compare the absolute figures against other reference values. This gives you a sense of the approximate orders of magnitude involved. For example, the entire German federal budget for 2016 provides for expenditure totalling EUR 316.9 billion

Roger Arnold's picture
Roger Arnold on Jul 4, 2017 5:46 am GMT

Bas, thank you for the link to Jacobson’s “debunking” of the paper by Clark et. al. It’s fascinating reading. But I’m afraid that, far from debunking Clark’s charges against his 100% renewables study, it deepens the hole that Jakobson has dug himself into.

The article you referenced is titled 30 False and 5 highly misleading statements in the main text of Clack et al.. It purports to be a point by point refutation of those statements. After reading it however, I came away convinced that the statements to which Jakobson is objecting are all substantially correct. His “refutations” are masterful exercises in misdirection and obfuscation. Jakobson is clearly a skilled manipulator and academic politician, but my opinion of him now is a good deal lower than it was before reading his defense.

As an example, let’s consider just the first “false” statement that Jakobson attempts to refute:

A large number of analyses and assessments, including those performed by the Intergovernmental Panel on Climate Change, the National Oceanic and Atmospheric Administration, the National Renewable Energy Laboratory, and the International Energy Agency have concluded that deployment of a diverse portfolio of clean energy technologies makes a transition to a low-carbon-emission energy system both more feasible and less costly.

How does Jakobson respond to that?

False. IPCC says the exact opposite: IPCC Working Group III, Chapter 7 Section P. 534.

…high shares of variable RE power, for example, may not be ideally complemented by nuclear, CCS, and CHP plants (without heat storage).

As it happens, the IPCC statement that Jakobson quotes is correct; in a scenario where variable renewables produce a high share of the supply, nuclear, CCS, and CHP plants without heat storage may — emphasis may — not be the ideal complements. The ideal complement for variable renewables, in a scenario that takes a high share for such renewables as a starting assumption, would have very low capital cost per kilowatt of capacity (since by stipulation it will operate with a low capacity factor, sitting idle most of the time), would support fast throttling over its entire power range with no loss of efficiency, and would dissipate no energy when standing by. Neither nuclear nor fossil fuels with CCS would appear to qualify. But Clark et. al did not claim that they did.

The statement that Clark et. al. make is specifically talking about the feasibility and cost of “a transition to a low-carbon-emission energy system”. And it’s 100% correct; a large number of analyses and assessments, including those performed by the IPCC, have concluded just that. Jakobson, however, chooses to take it as a statement about the ideal complement for variable renewables in a high penetration scenario.

Refuting a statement that you opponent didn’t make and claiming thereby to have refuted what he did say is a classic bit of rhetorical legerdemain beloved of political hacks. But it certainly ain’t science.

Roger Arnold's picture
Roger Arnold on Jul 4, 2017 6:50 am GMT

An interesting point I omitted from my preceding comment is that even in a scenario where a high share from variable renewables is stipulated, existing nuclear can be transformed into a nearly ideal complement for variable renewables.

All that’s needed is to parallel the NPP with a large discretionary load (or set of loads). Water desalination and production of electrolytic hydrogen and oxygen are good candidates. At times when variable renewables are serving 100% of the regular load, 100% of the NPP output goes to service the discretionary load. As the variable renewables fall off, power exported from the NPP is ramped up by ramping down what’s delivered to the discretionary load.

When an existing NPP is used in that manner, the capital cost of backing capacity is essentially zero; the capital cost of the NPP is sunk. There is, of course, capital expenditure for implementing the discretionary load(s), but since those loads are productive, the expenditure may be justified on its own merits.

One objection to this approach is that be baseload plant is “stealing” discretionary load from the variable renewables. Which need all the discretionary loads they can get. But discretionary loads alone are inadequate complements for variable renewables. They can be ramped up in response to excess supply and ramped down for declining supply. However they can’t be ramped down below zero when the supply from variable renewables falls below requirements for non-discretionary loads.

Darius Bentvels's picture
Darius Bentvels on Jul 4, 2017 9:48 am GMT

With more wind & solar, base load plants are gradually replaced by more flexible fossil plants (and of course other such as hydro & pumped storage).

When wind & solar increase further those flexible fossil plants are superseded by fast starting gas-turbines helped by batteries (and of course hydro & pumped storage) to cover the first minutes. It’s the phase Germany is now entering.

Then Power-to-Gas creating renewable gas (to store or use for other purposes such as car fuel) to make wind & solar overproduction useful (marginal costs of that electricity is near zero) is really taking off.
These will become substantial after ~2030 in Germany when wind+solar share surpasses 50% of production.

All in all this implies that CO2 emissions per generated KWh will continue to fall.

Nuclear power plants are either not flexible enough, or only flexible enough against extreme high costs, to be competitive in such environment.

Darius Bentvels's picture
Darius Bentvels on Jul 4, 2017 10:28 am GMT

Thank you for your thoughtful response.
Near all such discussions, as now between Jacobson and his opponents, are not science. If it would be science the matter would be solved by research results.
His opponents also refer to political statements such as the IPCC (their citation is questionable).

Essentially it’s an optimization issue:
With which means, and when used, is it cheapest to reach a 100% clean (=virtually no CO2 emitting) energy production.

With the present high costs and long implementation of nuclear, it’s clear that scenario’s including substantial new nuclear cannot compete against those without.

Even French govt institute ADEME concluded that in its scenario studies regarding the best solution for 2050.
Those results are behind the French drive to reduce nuclear fast (even faster than Germany).

Darius Bentvels's picture
Darius Bentvels on Jul 4, 2017 10:58 am GMT

That’s an interesting idea. However I’m afraid that the economics don’t work.
When wind & solar produce only 50%, they will already produce >100% during 25% of the time (Danish authorities).

That implies that the price will be <1cnt/KWh during 25% of the time. So the water desalination, hydrogen, etc. producing companies won't pay more…

Stronger, German aluminum smelters (which need lot of electricity when producing) made arrangements with their workforce that they only work when the weather forecast predicts high production. hence very low electricity prices.
So they flourish, while even US smelters have an hard time despite the lower wages…

The produced renewable gas can easily be stored in deep earth cavities and used when there is a period with little sun & wind. Experts expect that the overall efficiency of that process will become ~40%.

So when the PtG plant buys at an av. 1cnt/KWh (between negative and ~2cnt/KWh), the resulting electricity will cost 2.5cnt/KWh + equipment costs = ~3cnt/KWh.
Note that in Germany all those plants are scheduled to be unmanned, automated and housed in 'sea-containers'.

I really don't see how nuclear can compete in such environment (~av. price 1cnt 25% of the time, 3cnt during the rest).

The solution which many pro-nuclear choose; "100% by renewable impossible", is debunked by a.o. the French and German studies.

Helmut Frik's picture
Helmut Frik on Jul 4, 2017 1:33 pm GMT

Well, in germany biomass (biogas) units are one way to complement wind and solar. They have been running baselode, but now have regulatory incentive , and a little financial incentive to add gas storage and motor capacity to run as peaker plants (which is technically a easy task for them) I have into the wind and solar data of 2016 for Europe (EU part) and found that the biomas energy of germany in 2016 was enough to smmoth the wind and solar output flat with a small amount of the hydropower of the alps and germany for whole europe as it is installed today. (so same output around the clock the whole year). OK, engine capacity would have to be quite big, and a bit of curtailment happens too. (3-5% as I remember) And wind power and solar power are very much concentrated still in small areas here in europe. Rising renewable power production in all european countries to german level would spread them nmuch more evenly, and make the task of smoothing remaining variances much more easy.
Ant there is more biomass, now used for building heating (wood) and to blend gasoline and diesel which could be used as backup in comming decades. (but not as baseload like Drax)

Helmut Frik's picture
Helmut Frik on Jul 10, 2017 9:39 am GMT

And if you read the document you find that
a) power generation sector is ahead of shedule, no problems with renewables anywhere in sight
b) building sector is slightly behind shedule, needind some more push to increase change rate
c) traffic sector is far behind shedule, here there is the major problem.

Engineer- Poet's picture
Engineer- Poet on Jul 10, 2017 1:56 pm GMT

If the stalling of the reduction in emissions from the electric power sector (a direct, predictable and predicted consequence of the first plank of the Energiewende) is consistent with the schedule, there is something very wrong with the schedule.

Helmut Frik's picture
Helmut Frik on Jul 10, 2017 3:33 pm GMT

Power sector is reducing CO2 emissions even while exporting nearly 10% of the power output additionally.
But CO2 emission reductions were mainly expected from traffic sector and building sector at this time.

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »