"Fraunhofer ISE was wrong" - © Auke Hoekstra
Author profile picture

A new study on the positive sides of hydrogen and diesel was presented by the Fraunhofer ISE Institute on 13-7-2019 and this blog is my rebuttal to that. Here I refer to a recent publication in order to show where they have gone wrong. I hold them in high esteem; yet their study provides a misrepresentation of information which H2 Mobility paid for. When I’m finished with amending the study, the graph will illustrate that diesel no longer sounds that good and hydrogen is clearly not any better than electric anymore either. And it’s not a case of ‘he-said-she-said’; you should be able to follow it all.

If you only want to know the highlights, just read the bold text.


“Fraunhofer ISE was wrong” – © Auke Hoekstra

The electric vehicle is a threat to the oil and car industry. Maybe that’s why H2 Mobility were happy to pay for a study with a range of assumptions which are unfavourable towards the electric vehicle.

The rise of the electric vehicle is good news to some but abhorrent to others, which is why I keep debunking studies (see debunk one], [two], [three] en [my reasons]) that psurport to show electric vehicles are not better for the environment. The latest one I came across was paid for by H2 Mobility – and if you are wondering who is paying them: this picture shows their main sponsors.

“Fraunhofer ISE was wrong” – © Auke Hoekstra

The study was done under the banner of the reputable Fraunhofer Institute. However, there are as many as 72 largely independent consultancy groups named Fraunhofer. Not all Fraunhofer experts were happy with the study, and some even applauded me on twitter.

The authors confronted my criticism in the German Handelsblatt newspaper. I did not get the opportunity to respond to that in the Handelsblatt, but I do refer to their response in this blog.

As you will see, we don’t need to delve into methodology as that is not where we differ from each other. It is primarily about assumptions. And these are often the most important. As Ronnie Belmans stated on twitter in reaction to my thread: “My old German prof. told me during my PhD: ‘Read the assumptions before reading conclusions. You avoid wasting time.’ ”

Or as one of my own professors used to say: “as-u-me” makes an “ass” out of “u” and “me”.

Hence, I’ll point out why I think some of their assumptions are wrong and how this invalidates their conclusion that hydrogen cars are better for the environment than electric vehicles.

One word in advance: hydrogen is much cleaner than gasoline or diesel and I think it’s going to be very important as a long-term storage medium. So no, I don’t hate hydrogen! And I’m not proposing that every household in the world buys a private electric vehicle with a large battery. That would cause enormous environmental damage and make cities less pleasant than they might otherwise be. I simply don’t approve of spin as a way to create FUD (fear uncertainty and doubt) around electric vehicles.

First things first

In a recent presentation, I outlined the *Top 6 Mistakes* which make electric vehicles look bad. Author André Sternberg et al. of the Fraunhofer ISE study (abbreviated as Sternberg) manage to make all 6 of those mistakes and add what I’d call a dirty trick.

The original Fraunhofer study can be found here: https://www.ise.fraunhofer.de/de/presse-und-medien/news/2019/fraunhofer-ise-vergleicht-treibhausgas-emissionen-von-batterie-und-brennstoffzellenfahrzeugen.html

In my critique, I’ll often refer to my article on the top 6 mistakes titled “The Underestimated Potential of Battery-driven Electric Vehicles to Reduce Emissions” and published in Joule, the online open access science archive. Therefore, if you really want to check everything out you should probably read that publication too. It is fairly brief.

“Fraunhofer ISE was wrong” – © Auke Hoekstra

As stated, Sternberg et al., have made 6 standard mistakes plus used a dubious strategem. The following chart provides a preview for you and offers an initial indication of the alterations that these rectifications achieve. As you can see, the corrections lead to a reduction in emissions in all cases, although they drop the least for diesel and the most for electric vehicles.

“Fraunhofer ISE was wrong” – © Auke Hoekstra

Let me now explain the rectifications step by step.

Fraunhofer proposes making hydrogen from wind and power batteries with solar energy. By assuming solar emits 3.5 times more CO2, they mask the energy losses from hydrogen production. I tally the same electricity for both.

A ‘dirty trick’ that proponents of hydrogen often use in order to make hydrogen sound good, is to assume the hydrogen is made from the cleanest sources possible and then compare it to BEVs running on a relatively dirty electricity mix. But that’s like comparing apples with oranges!

In this study, we discover that trick along with a clever twist. Hydrogen and batteries both use clean energy: hydrogen is made from wind and batteries use solar energy. All good right?

Wrong, because if you look closely, Sternberg assumes that wind energy emits only 14 grams of CO2 equivalent to every kWh that is produced. Whereas solar emits 48 grams, or 3.5 times as much. That way they can obscure the fact that producing hydrogen costs 3 times as much energy as when using a battery directly. This is because you lose energy when you convert electricity to hydrogen, you lose energy when you compress and clean the hydrogen and you lose energy when you burn the hydrogen in the fuel cell.

After my criticism, the authors defended their argument by saying that they assumed ocean winds would make hydrogen at sea directly, while BEVs would be able to drive using the electricity from their solar rooves. And this ties into the narrative: “but electricity transportation is so hard whereas hydrogen transportation is so easy”. It is true that at really large volumes, a pipeline with hydrogen transports energy cheaper than an electricity cable does. But so far, building windmills and making hydrogen are much more expensive steps in the process, and if you were to use windmills for BEV’s, you could drive three times as many cars with just one windmill!

Quite apart from the fact that I make energy models – a battery in a car is the ideal way to absorb excess wind energy at night and excess solar energy during the day. Consequently, not using the battery to store night-time wind is not just arbitrary but also expensive. When it comes to a system, this is all highly questionable. So that’s when I cry foul play, and then make them all run on the same type of renewable energy. Doesn’t really matter which one but I’ve opted for solar energy.

Although one thing that I have let slide is that they take an absurdly economical diesel hybrid engine from a Hyundai Tucson 1.6 CRDi and did not add the energy needed to produce this engine nor its battery. The car is so new that I couldn’t find a road test in order to check against the often far too ‘carmaker-friendly’ WLTP, and there is (still?) not a more reliable EPA rating for this car.

Two minor tweaks replace a somewhat less energy-intensive chassis (for all cars) and use standard IPCC values for diesel emissions. Basically, because I have good sources for that (see publication) and because they haven’t mentioned theirs. Nonetheless, these two tweaks only make a modest difference.


“Fraunhofer ISE was wrong” – © Auke Hoekstra

Fraunhofer claims there is no data that indicates that batteries last longer than 150 thousand km. I provide an abundance of sources that show that a battery will easily last 600 thousand km. Even when export and second lives have been ignored, 300 thousand km (similar to a diesel) seems a bit conservative.

An assumption that plagues many car studies is that cars are scrapped after 150 thousand kilometres. Maybe this is because consultants often drive new lease vehicles. But it is not correct. Good data on how long cars last is hard to find, however, Gneiwomir Flis helped me out tremendously last week and as you can see in the resulting graph, the average lifespan is around ten years (eight in Germany). Now if sales and scrappage were constant, the age at which vehicles are scrapped would be double the average age. So, 16 in Germany and 20 in Europe in general. We also know cars drive around 14 thousand km a year in Germany which would mean 16 x 14 = 224 thousand km before being scrapped.


“Fraunhofer ISE was wrong” – © Auke Hoekstra

More research done by Flis (see aforementioned thread) revealed that cars in the UK are scrapped after 15 years of use and in Norway after around 17 years. Accordingly, 15 or 16 years for Germany might not be such a strange assumption. But there are four things that have not been taken into account here.

Firstly, cars from rich countries like Germany often have a second life in poorer countries like Poland. Which is why the average age is 8 years in Germany and 18 years in Poland. One way to make these two numbers work is to assume that some cars in Poland come from Germany after let’s say 10 years and stay on the road until they are scrapped at let’s say 25 years. Consequently, in a life cycle analysis, we would have to add a couple of years to the age of the German car in order to get a good idea of when the car and battery have really reached the end of their life.

Secondly, cars are mostly scrapped when the engine becomes old and engine upkeep becomes too expensive. This is one reason diesel cars are popular: their engines last longer. According to Buchal et al., (see my publication) a diesel lasts around 300 thousand kilometres. Logically speaking, that would mean that electric cars last even longer because according to recent studies of present-day batteries, they last anywhere between 1300 (most pessimistic) and 10 000 (optimistic) cycles. A battery of 90 kWh has a range of 450 km, so 1300 cycles is 1300 x 450 = 585 thousand kilometres. And that’s the absolute worst-case scenario. Let’s make that 600 thousand for now.

Thirdly, cars that are cheaper per km get driven more km. That is not so strange of course. It’s the reason that diesel and LPG (with their lower taxes) are driven more on average than gasoline vehicles. Based on that, we would expect electric vehicles to be driven more as well as they are much cheaper per km than diesel or LPG.

Fourthly, batteries can have a second life or be recycled. After 600 thousand kilometres the battery still has a capacity of 80%.

Co-author Christoph Hank’s defence is that there are only “incidental reports” of batteries lasting very long. That is simply not true. My friend Prof. Steinbuch states in his blogpost that documents these values from hundreds of Tesla drivers gathered by Merijn Couman: “Batteries have on average 91% remaining after 270 000 km. If that linear behaviour were to continue, then the ‘lifespan’ (with 80% capacity still left) would be 820 000 km.” Additional 2018 sources here, here, here, here, here, here, here and here.

For all of these reasons, I think that 150 thousand km is unlikely. Sixteen years and 14 thousand kilometres per year already brings you up to 224 thousand km for an average car in Germany. Add the fact that the drivetrain, battery and brakes are still basically brand new at that moment, and that the car is extremely cheap to drive per km, then I think to tally as many km as a diesel is wildly conservative. Which is why I take the figure of 300 thousand kilometres. And that still excludes the second life of the car in Poland and the second life of the battery, ideally both of which you should include in the LCA.

Using recent more detailed data, battery production emits around 65 kg/kWh, not 133 kg/kWh.

Fraunhofer assumes battery production to be 133 kg/kWh, based on a very dirty electricity mix (805 gr CO2/kWh). Authors Hebling and Hank claim that most batteries come from Asia and hence they should take that dirty mix into account. But the elephant in the room is that Tesla is the big competitor in Germany and their cars are not made in Asia at all. Tesla even claims that its factory runs entirely on renewable energy. So, they should at least have taken a certain weighted average into account. Plus, if they talk about German or European cars, the average should go down because of Tesla.

When it comes to batteries they use their own observation that a battery pack weighs in at 7.5 kg/kWh (= 135 Wh/kg) while a Tesla Model 3 is already at 6.25 kg/kWh (= 160Wh/kg) and it’s not even 2020 yet. Again, that’s 19% better.

All this despite the fact that that they have rolled out their own calculations based on an outdated scientific publication from 2014. Whereas I use scientific publications from 2017 and 2019 aimed specifically at these figures which include current production in Asia (see ref 6 and 7). We especially see that cell assembly has become much more energy efficient in larger factories. Authors Hebling and Hank’s defence is that part of those references were not yet available when they wrote the study.

All in all, battery production is quickly become more energy efficient, also because batteries require increasingly less material all the time nowadays. E.g. new Tesla batteries contain a third of cobalt compared to previous versions, and no cobalt even seems to be feasible. Hence the data I use now will undoubtedly be too pessimistic by the time it’s 2020.

The chart shows the results after remedying those two assumptions.

“Fraunhofer ISE was wrong” – © Auke Hoekstra

The report takes the dirty German mix into account yet doesn’t talk about hydrogen vs electric batteries in Germany at all. Therefore, I take the European mix (pretty close to the US mix by the way) so as to make the study more representative.

Finally, let’s look at the average electricity mix that is assumed for electric vehicles and emits (you guessed it!) way too much CO2. As I said before it’s a common trick: proponents of hydrogen never make comparisons with countries who use a clean energy mix, because it makes hydrogen from natural gas look especially bad (currently more than 90% of hydrogen comes from natural gas).

The German mix is well suited to their needs as it is currently relatively dirty (after the ‘Atomausstieg’ – phasing out of nuclear energy -that made coal dominant). Yet there are plans and even laws that stipulate that they must clean up the mix rather quickly (as part of the ‘Kohlenausstieg’ – phasing out of coal). However, this process is always hard to predict for a country. It is also less relevant as nobody talks about hydrogen vs battery-driven electric in Germany. It is not addressed in the report (unless you read the data values in the annexe) and not covered in the press.

So as to make the report more relevant I took into account the EU electricity values over the lifetime of an electric vehicle. Additionally, the trend in a CO2 reduction of the mix per year is pretty constant for the EU as a whole. Which makes it possible, using the average energy mix, to estimate each year how much cleaner e-driving is becoming.

The graph shows the effect of all the rectifications that have been made.


“Fraunhofer ISE was wrong” – © Auke Hoekstra

The end result: even the most economical diesel is no match for hydrogen but driving battery-driven electric vehicles emits even less CO2.

Some final notes before people say that I am pretending that electric vehicles are unproblematic. I’m saying they are much better than diesel and slightly better than hydrogen – nevertheless, just switching over to battery-driven electric cars is simply not enough!

Of course, using a bicycle or walking is better for your health and the planet than driving a car is.

We must also try to make the entire supply chain (including mining operations) run on renewable electricity. And of course, we must recycle.

Furthermore, I believe that in the future we will see lots of shared autonomous vehicles that will be optimized for individual trips. That would mean two things: roughly ten times fewer vehicles on the roads, and – since we make most trips alone and over short distances – the vehicle size would also become much smaller. This is what I call public transit 2.0 and that, together with a supply chain made up of renewable energy, would slash vehicle production emissions by a factor of ten or more. Emissions while driving would be reduced 2 – 4 x because of the smaller and lighter vehicles. That is the future I am working towards. Not one in which everybody in the world has their own electric car.


“Fraunhofer ISE was wrong” – © Auke Hoekstra

On the same subject:
Here is part 1 of our mini series “Mobility of the Future – Battery or hydrogen or both?”
Here is part 2 of our mini series “Mobility of the Future – Battery or hydrogen or both?”
Here is part 3 of tour mini series “Mobility of the Future – Battery or hydrogen or both?”