Another study on the ERoEI of solar PV

10 05 2016

Originally posted by Euan Mearns on his blog Energy Matters, this study makes Pedro Prieto’s look very good….. the differences in ERoEI between the two studies must be a function of the difference in latitude between Spain and the UK, and even possibly by the fact that as the ERoEI of fossil fuels drops like a stone, the ERoEI of renewables must follow suit, as they rely entirely on the former.  Is Mearns a fan of nuclear power?  You make up your mind….

As Fort McMurray burns, and its smoke plume reaches the East coast of the USA, it’s occurred to me that the inevitable efforts and energy smoke-plume-from-fort-mcmurray-fire-reaches-us-east-coastrequired to rebuild it once the fire is out (IF, that is, it doesn’t reach the tar sands and sets them alight…), should be included in the ERoEI of fossil fuels.  Whilst it’s impossible to say Climate Change caused the fire in Alberta Canada, it’s impossible to not make the connection that the only reason it was over 20°C when the fire started was entirely down to the jetstream going haywire because of the arctic melt……  in fact, the energy spent rebuilding destroyed infrastructure caused by Climate Change anywhere should now be included in the ERoEI of fossil fuels….



A new study by Ferroni and Hopkirk [1] estimates the ERoEI of temperate latitude solar photovoltaic (PV) systems to be 0.83. If correct, that means more energy is used to make the PV panels than will ever be recovered from them during their 25 year lifetime. A PV panel will produce more CO2 than if coal were simply used directly to make electricity. Worse than that, all the CO2 from PV production is in the atmosphere today, while burning coal to make electricity, the emissions would be spread over the 25 year period. The image shows the true green credentials of solar PV where industrial wastelands have been created in China so that Europeans can make believe they are reducing CO2 emissions (image credit Business Insider).

I have been asked to write a post reviewing the concept of energy return on energy invested (ER0EI) and as a first step in that direction I sent an email to my State-side friends Charlie Hall, Nate Hagens and David Murphy asking that they send me recent literature. The first paper I read was by Ferruccio Ferroni and Robert J. Hopkirk titled Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation [1] and the findings are so stunning that I felt compelled to write this post immediately.

So what is ERoEI? It is simply the ratio of energy gathered to the amount of energy used to gather the energy (the energy invested):

ERoEI = energy gathered / energy invested

Simple, isn’t it? Well it’s not quite so simple as it appears at first sight. For example, using PV to illustrate the point, the energy gathered will depend on latitude, the amount of sunshine, the orientation of the panels and also on the lifetime of the panels themselves. And how do you record or measure the energy invested? Do you simply measure the electricity used at the PV factory, or do you include the energy consumed by the workers and the miners who mined the silicon and the coal that is used to make the electricity? Ferroni and Hopkirk go into all of these details and come up with an ERoEI for temperate latitude solar PV of 0.83. At this level, solar PV is not an energy source but is an energy sink. That is for Switzerland and Germany. It will be much worse in Aberdeen!

Why is ERoEI important? It is a concept that is alien to most individuals, including many engineers, energy sector employees, academics and policy makers. The related concept of net energy is defined as:

Net Energy = ERoEI – 1 (where 1 is the energy invested)

Net energy is the surplus energy left over from our energy gathering activities that is used to power society – build hospitals, schools, aircraft carriers and to grow food. In the past the ERoEI of our primary energy sources – oil, gas and coal – was so high, probably over 50, that there was bucket loads of cheap energy left over to build all the infrastructure and to feed all the people that now inhabit The Earth. But with the net energy equation for solar PV looking like this:

0.83-1 = -0.17

….. Brussels we have a problem!

So how can it be possible that we are managing to deploy devices that evidently consume rather than produce energy? The simple answer is that our finance system, laws and subsidies are able to bend the laws of physics and thermodynamics for so long as we have enough high ERoEI energy available to maintain the whole system and to subsidise parasitic renewables. Try mining and purifying silicon using an electric mining machine powered by The Sun and the laws of physics will re-establish themselves quite quickly.

In very simple terms, solar PV deployed in northern Europe can be viewed as coal burned in China used to generate electricity over here. All of the CO2 emissions, that underpin the motive for PV, are made in China. Only in the event of high energy gain in the PV device would solar PV reduce CO2 emissions. More on that later.

Energy Return

The calculations are all based on the energy produced by 1 m^2 of PV.

Theoretical calculations of what PV modules should generate made by manufacturers do not take into account operational degradation due to surface dirt. Nor do they take into account poor orientation, unit failure or breakage, all of which are quite common.

The actual energy produced using Swiss statistics works out at 106kWe/m^2 yr

We then also need to know how long the panels last. Manufacturers claim 30 years while empirical evidence suggests a mean scrapage age of only 17 years in Germany. Ferroni and Hopkirk use a generous 25 year unit life.

Combining all these factors leads to a number of 2203kWe/m^2 for the life of a unit.

Energy Invested

The energy invested calculation is also based on 1 m^2 of panel and uses mass of materials as a proxy for energy consumed and GDP energy intensity as a proxy for the labour part of the equation.

Two different methods for measuring energy invested are described:

  • ERoEI(IEA)
  • ERoEI(Ext)

Where IEA = methodology employed by the International Energy Agency and Ext = extended boundary as described by Murphy and Hall, 2010 [2,3]. The difference between the two is that the IEA is tending to focus on the energy used in the factory process while the extended methodology of Murphy and Hall, 2010 includes activities such as mining, purifying and transporting the silicon raw material.

In my opinion, Ferroni and Hopkirk correctly follow the extended ERoEI methodology of Murphy and Hall and include the following in their calculations:

  • Materials to make panels but also to erect and install panels
  • Labour at every stage of the process from mining manufacture and disposal
  • Manufacturing process i.e. the energy used in the various factories
  • Faulty panels that are discarded
  • Capital which is viewed as the utilisation of pre-existing infrastructure and energy investment
  • Integration of intermittent PV onto the grid

And that gives us the result of ERoEI:

2203 / 2664 kW he/m^2 = 0.83

The only point I would question is the inclusion of the energy cost of capital. All energy produced can be divided into energy used to gather energy and energy for society and I would question whether the cost of capital does not fall into the latter category?

But there appears to be one major omission and that is the energy cost of distribution. In Europe, about 50% of the cost of electricity (excluding taxes) falls to the grid construction and maintenance. If that was to be included it would make another serious dent in the ERoEI.

This value for ERoEI is lower than the value of 2 reported by Prieto and Hall [4] and substantially lower that the values of 5 to 6 reported by the IEA [5]. One reason for this is that the current paper [1] is specifically for temperate latitude solar. But Ferroni and Hopkirk also detail omissions by the IEA as summarised below.

IEA energy input omissions and errors

a) The energy flux across the system boundaries and invested for the labour is not included.
b) The energy flux across the system boundaries and invested for the capital is not included.
c) The energy invested for integration of the PV-generated electricity into a complex and flexible electricity supply and distribution system is not included (energy production does not follow the needs of the customer).
d) The IEA guidelines specify the use of “primary energy equivalent” as a basis. However, since the energy returned is measured as secondary electrical energy, the energy carrier itself, and since some 64% to 67% of the energy invested for the production of solar-silicon and PV modules is also in the form of electricity (Weissbach et al., 2013) and since moreover, the rules for the conversion from carrier or secondary energy back to primary energy are not scientifically perfect (Giampietro and Sorman, 2013), it is both easier and more appropriate to express the energy invested as electrical energy. The direct contribution of fossil fuel, for instance in providing energy for process heating, also has to be converted into secondary energy. The conversion from a fossil fuel’s internal chemical energy to electricity is achieved in modern power plants with an efficiency of 38% according to the BP statistic protocol (BP Statistical Review of World Energy, June 2015). In the present paper, in order to avoid conversion errors, we shall continue to use electrical (i.e. secondary) energy in kW he/m2 as our basic energy unit.
e) The recommended plant lifetime of 30 years, based on the experiences to date, must be regarded as unrealistic.
f) The energy returned can and should be based on actual experimental data measured in the field. Use of this procedure will yield values in general much lower than the electricity production expected by investors and politicians.

Of those I’d agree straight off with “a”, “c” and “f”. I’m not sure about “b” and “e” I’m sure this will be subject to debate. “d” is a complex issue and is in fact the same one described in my recent post EU and BP Renewable Electricity Accounting Methodologies. I agree with Ferroni and Hopkirk that units of electricity should be used throughout but if the IEA have grossed up the electricity used to account for thermal losses in power stations then this would increase their energy invested and suppress not inflate their estimates of ERoEI. Hence this is a point that needs to be clarified.

Environmental impacts

The main reason for deploying solar PV in Europe is to lower CO2 emissions. The European Commission and most European governments have been living in cloud cuckoo land allowing CO2 intensive industries to move to China, lowering emissions in Europe while raising emissions in China and making believe that importing steel from China somehow is emissions free.

The example of solar PV brings this into sharp focus. Assuming the main energy input is from coal (and low efficiency dirty coal at that) and with ERoEI <1, making electricity from solar PV will actually create higher emissions than had coal been used directly to make electricity for consumption in the first place. But it’s a lot worse than that. All of the emissions associated with 25 years of electricity production are in the atmosphere now making global warming much worse than it would otherwise have been without the PV.

And it gets even worse than that! The manufacture of PV panels involves lots of nasty chemicals too:

Many potentially hazardous chemicals are used during the production of solar modules. To be mentioned here is, for instance, nitrogen trifluoride (NF3), (Arnold et al., 2013), a gas used for the cleaning of the remaining silicon-containing contaminants in process chambers. According to the IPCC (Intergovernmental Panel on Climate Change) this gas has a global warming potential of approximately 16600 times that of CO2. Two other similarly undesirable “greenhouse” gases appearing are hexafluoroethane (C2F6) and sulphur hexafluoride (SF6).


The average weight of a photovoltaic module is 16 kg/m2 and the weight of the support system, inverter and the balance of the system is at least 25 kg/m2 (Myrans, 2009), whereby the weight of concrete is not included. Also, most chemicals used, such as acids/ bases, etchants, elemental gases, dopants, photolithographic chemicals etc. are not included, since quantities are small. But, we must add hydrochloric acid (HCl): the production of the solar- grade silicon for one square meter of panel area requires 3.5 kg of concentrated hydrochloric acid.

Comparison with nuclear

The paper offers some interesting comparisons with nuclear power. Looking first at materials used per unit of electricity produced:

  • PV uses 20.2 g per kW he (mainly steel aluminium and copper)
  • A nuclear power station uses 0.31 g per kW he (mainly steel) for a load factor of 85%

kW he = kilowatt hours electrical

Looking at labour, the authors observe:

The suppliers involved in the renewable energies industry advertise their capability to create many new jobs.

While of course the best forms of energy use as little labour as possible. At the point where ERoEI reaches 1, everyone is engaged in gathering energy and society as we know it collapses!

  • Solar PV creates 94.4 jobs per MW installed, adjusted for capacity factor.
  • Nuclear creates 13 jobs per MW installed covering construction, operation and decommissioning.

This may seem great to the politicians but it’s this inefficiency that makes solar PV expensive and kills the ERoEI. And looking at capital costs:

  • Solar PV needs CHF 6000 per kW installed (CHF = Swiss Franc)
  • Nuclear power CHF 5500 per kW installed

But normalising for capacity factors of 9% for solar and 85% for nuclear we get for effective capacity:

66,667 / 6471 = 10.3

Solar PV is 10 times more capital intensive than nuclear.

Energy transformation

When ERoEI approaches or goes below 1 we enter the realm of energy transformation which is quite common in our energy system. For example, converting coal to electricity we lose approximately 62% of the thermal energy. Converting coal and other raw materials into a PV panel may in certain circumstances make some sense. For example PV and a battery system may provide African villages with some electricity where there is little hope of ever getting a grid connection. Likewise for a mountain cabin. Individuals concerned about blackouts may also consider a PV battery system as a backup contingency.

But as a means of reducing CO2 emissions PV fails the test badly at temperate latitudes. It simply adds cost and noise to the system. In sunnier climates the situation will improve.

Concluding comments

The findings of this single study suggest that deploying solar PV at high latitudes in countries like Germany and the UK is a total waste of time, energy and money. All that is achieved is to raise the price of electricity and destabilise the grid. Defenders of RE and solar will point out that this is a single paper and there are certainly some of the inputs to Ferroni and Hopkirk that are open to debate. But there are reasons to believe that the findings are zeroing in on reality. For example Prieto and Hall found ERoEI for solar PV = 2. Looking only at cloudy, high temperate latitudes will substantially degrade that number.

And you just need to look at the outputs as shown below. Solar PV produces a dribble in winter and absolutely nothing at the 18:00 peak demand. There is a large financial cost and energy cost to compensate for this that RE enthusiasts dismiss with a wave of the arm.

Figure 1 From UK Grid Graphed. The distribution of solar production in the UK has grown 7 fold in 4 years. But 7 times a dribble in winter is still a dribble.  The large amount of embodied energy in these expensive devices does no work for us at all when we need it most.

Energy Matters has a good search facility top right. Insert solar pv and I was surprised to find how many articles Roger and I have written and they all more or less reach the same conclusions. I have added these links at the end of the post.

Figure 2 A typical solar installation in Aberdeen where the panels are on an east facing roof leaving the ideal south facing roof empty. This is a symbol of ignorance and stupidity that also pervades academia. Has anyone seen a University that does not have solar PV deployed? I’ve heard academics argue that orientation does not matter in Scotland, and they could be right. I dare say leaving the panels in their box would make little difference to their output. Academics, of course, are increasingly keen to support government policies. Note that sunny days like this one are extremely rare in Aberdeen. And in winter time, the sun rises about 10:00 and sets around 15:00.

Two years ago I fulminated about the random orientation of solar panels in Aberdeen in a post called Solar Scotland. And this random orientation will undoubtedly lead to serious degradation of the ERoEI. PV enthusiasts will no doubt assume that all solar PV panels are optimally orientated in their net energy analysis while in the real world of Ferroni and Hopkirk, they are not. A good remedy here would be to remove the feed in tariffs of systems not optimally deployed while ending future solar PV feed in tariffs all together.

But how to get this message heard at the political level? David MacKay’s final interview was very revealing:

The only reason solar got on the table was democracy. The MPs wanted to have a solar feed-in-tariff. So in spite of the civil servants advising ministers, ‘no, we shouldn’t subsidise solar’, we ended up having this policy. There was very successful lobbying by the solar lobbyists as well. So now there’s this widespread belief that solar is a wonderful thing, even though … Britain is one of the darkest countries in the world.

If the politicians do not now listen to the advice of one of the World’s most famous and respected energy analysts then I guess they will not listen to anyone. But they will with time become increasingly aware of the consequences of leading their electorate off the net energy cliff.


[1] Ferruccio Ferroni and Robert J. Hopkirk 2016: Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation: Energy Policy 94 (2016) 336–344

[2] Murphy, D.J.R., Hall, C.A.S., 2010. Year in review-EROI or energy return on (energy) invested. Ann. N. Y. Acad. Sci. Spec. Issue Ecol. Econ. Rev. 1185, 102–118.

[3] Murphy, D.J.R., Hall, C.A.S., 2011. Energy return on investment, peak oil and the end of economic growth. Ann. N.Y. Acad. Sci. Spec. Issue Ecol. Econ. 1219, 52–72.

[4] Prieto, P.A., Hall, C.A.S., 2013. Spain’s Photovoltaic Revolution – The Energy Return on Investment. By Pedro A. Prieto and Charles A.S. Hall, Springer.

[5] IEA-PVPS T12, Methodology Guidelines on the Life Cycle Assessment of Photovoltaic Electricity – Report IEA-PVPS T12-03:2011.




5 responses

10 05 2016
Chris Harries

If the interest here is to give nuclear energy a leg up, by comparing it favourably to solar, I would like to see a more comprehensive calculation of nuclear ER0EI. The industry itself puts out fabulously high figures, whereas academically nuclear ERoEI is normally reported at about the 5:1 level. What happens if true cradle-to-grave ERoEI is carried out nuclear?

Nuclear advocates do run their optimistic calculations on mooted new breeder technologies and it may be true that ERoEI for these, if they ever see the light of day, could be significantly higher than traditional nuclear.

All in all, I get rather annoyed when ERoEI gets used to pit one energy generator against another because the figures tend to be distorted by whichever supply-side advocate is running the story, and the focus tends to be along the lines of our dominant cultural paradigm: ‘How should we power industrial civilisation?’

In practice, comparing nuclear to wind energy would probably be more salient to that debate, since they both operate as centralised energy sources feeding into grids. Large scale centralised solar has been a disappointing failure to date.

11 05 2016

I find his adherence to nuclear somewhat misplaced!, IMHO! Nuclear extremely expensive, especially when you consider waste management!

13 05 2016
Chris Harries

Here’s an opposite take on centralised solar. The proof of the pudding will be in the eating. Am happy to be surprised, but so far centralised solar –economics and energy economics – haven’t worked out so well:

18 02 2017

Hmm. I think one has to be careful here. There is huge variation in what various academic papers report is the correct EROI of solar PV, and as ever it’s easy to “cherry pick” your references in order to prove or disprove a point. Ferroni and Hopkirk’s conclusions have been comprehensively debunked in another paper published in the same journal:

18 02 2017

” Ferroni and Hopkirk adopt ‘extended’ boundaries for their analysis of PV without acknowledging that such choice of boundaries makes their results incompatible with those for all other technologies that have been analysed using more conventional boundaries”

Charlie Hall speaks of this in a podcast I downloaded some time ago.

He calls it “mental masturbation”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s