Via Low-tech Magazine, 03-03-2008

New research shows, albeit unintentional, that generating electricity with solar panels can also be a very bad idea. In some cases, producing electricity by solar panels releases more greenhouse gases than producing electricity by gas or even coal.

Producing electricity from solar cells reduces air pollutants and greenhouse gases by about 90 percent in comparison to using conventional fossil fuel technologies, claims a study to be published this month in “Environmental Science & Technology”. Good news, it seems, until one reads the report itself. The researchers come up with a solid set of figures. However, they interpret them in a rather optimistic way. Some recalculations (skip this article if you get annoyed by numbers) produce striking conclusions.


Solar panels don’t come falling out of the sky – they have to be manufactured. Similar to computer chips, this is a dirty and energy-intensive process. First, raw materials have to be mined: quartz sand for silicon cells, metal ore for thin film cells. Next, these materials have to be treated, following different steps (in the case of silicon cells these are purification, crystallization and wafering). Finally, these upgraded materials have to be manufactured into solar cells, and assembled into modules. All these processes produce air pollution and heavy metal emissions, and they consume energy – which brings about more air pollution, heavy metal emissions and also greenhouse gases.

Energy mix

The ecological burden of energy use depends on the way electricity was generated. Therefore, the researchers bring into account 3 scenarios. One is based on the average European energy mix, another on the average American energy mix (which is about 45% more CO2-intensive). A third scenario uses the figures of the recent “CrystalClear” European Commission project, which investigated the real energy mix used by 11 European and American silicon and PV module manufacturing factories. Since they use comparatively more gas and hydropower, this is the best case scenario. The researchers investigated 4 types of solar cells: multi-crystalline silicon (with an efficiency of 13%), mono-crystalline silicon (14%), ribbon silicon (11.5%), and thin-film cadmium telluride (9%).

The optimistic conclusions of the researchers are based on a life expectancy of 30 years and solar insolation in the Mediterranean

The scientists come up with figures concerning the amount of greenhouses gasses emitted per kilowatt-hour of electricity delivered by one square meter of solar cells. They do that for every type of cell and for the three different scenarios. Thin film solar cells get the best score with 20.5 grams of CO2 in the European energy mix and 25 grams of CO2 in the American energy mix. In spite of their lower efficiency, they are more eco-friendly because they need less material and no aluminium frame. In spite of their high efficiency, mono-crystalline silicon cells score worst, with 43 grams of CO2 in the EU, and 55 gram of CO2-equivalent in the US. All other types and scenarios fit between these two extremes.

Solar insolation

However, these conclusions are dependent on some assumptions, most importantly solar insolation (the amount of sunlight that the cells receive) and lifetime expectancy. For solar insolation, the researchers choose 1,700 kWh per m² per year, which is the average of sunlight in Southern Europe. For lifetime expectancy, they choose 30 years. From these variables, they calculate the total lifetime electricity generation of one square meter of solar cells. Next, they divide the amount of CO2 emitted for the production of one square meter of solar panels by this lifetime electricity generation – and that’s how they achieve their conclusions.

Surprisingly, the key data of the calculation (the amount of CO2 emitted per square meter of solar panels) are nowhere to found in the report

Surprisingly, the key data of the calculation (the amount of CO2 emitted per square meter of solar panels) are nowhere to found in the report. That’s remarkable, since these data are the most objective numbers available. Even so, they can be calculated by multiplying the obtained results (in gram CO2 emitted per kilowatt-hour of generated electricity) by the lifetime electricity generation. This calculation gives the amount of greenhouse gases emitted for the production of one square meter of solar panels, regardless of the assumptions on solar insolation and lifetime expectancy.

2 to 20 flights

Once calculated, it’s not so suprising that the researchers choose not to write these figures down. In the best case scenario, one square meter of solar cells carries a burden of 75 kilograms of CO2. In the worst case scenario, that becomes 314 kilograms of CO2. With a solar insolation of 1,700 kWh/m²/yr an average household needs 8 to 10 square meters of solar panels, with a solar insolation of 900 kWh/m²/yr this becomes 16 to 20 square meters. Which means that the total CO2 debt of a solar installation is 600 to 3,140 kilograms of CO2 in sunny places, and 1,200 to 6,280 kilograms of CO2 in less sunny regions.  These numbers equate to 2 to 20 flights Brussels-Lissabon (up and down, per passenger) – source CO2 emissions Boeing 747.

According to the researchers, producing the same amount of electricity by fossil fuel generates at least 10 times as much greenhouse gasses. Checking different sources, this claim is confirmed: 1 kilowatt-hour of electricity generated by fossil fuels indeed emits 10 times as much CO2 (around 450 grams of CO2 per kWh for gas and 850 for coal). Solar panels might be far from an ideal solution, but they are definitely a better choice compared to electricity generated by fossil fuels. At least if we follow the assumptions chosen by the researchers.


Logically, if we make the same calculations for a solar insolation of 900 kWh/m² (the yearly average in Western Europe and in the Northeast and Northwest USA), the results get worse. In the worst case scenario (US grid, mono-crystalline silicon), emissions rise to 104 gram CO2 per kilowatt-hour of solar generated electricity, which makes solar panels only 4 times cleaner than gas. Now let’s play a bit with the life expectancy.

If we combine this lower solar insolation with an expected lifetime of only 15 years, the worst case scenario becomes 207 grams of CO2 per kilowatt-hour – just 2 times better than gas. Agreed, this is the worst case scenario, and even in that case solar panels are still a better choice than fossil fuels. But it becomes quite hard to describe them as a “clean” source of fuel.

A better strategy would be to use already available solar panels to produce more solar panels.

The life expectancy chosen by the researchers is, well, just an expectation. It’s true that most manufacturers give warranties of 20 to 25 years, so technologically speaking a life expectancy of 30 years is not implausible. However, there are other than technological reasons that may lead to a significantly lower life expectancy. The scientists note that the environmental score of solar panels will improve, because they are becoming more efficient each year. (They also become thinner, so less energy is needed to make them). Most likely they will also become cheaper.

Life expectancy

That means that in 15 to 20 years time, if the evolution in efficiency carries on the way it does now, a solar panel with an efficiency of 10 percent produced today will have to compete with cheaper solar panels that have efficiencies of about 20 percent. Moreover, and that’s a fact that the researchers are not taking into account, solar cells degrade in time. Typically, the warranty given by solar cell manufacturers covers just 80 percent of power output. All this means that it may make economic sense to substitute older panels with newer panels before they are 30 years old. Again, even in that case the ecological score will probably still be better than the one of fossil fuels, but the point is that the gap can become very small.

For rooftop and ground-base installations, the eco-friendliness can be good or doubtful, depending on the solar insolation and the life expectancy. But if we consider solar panels mounted on gadgets like laptops or mobile phones, solar energy becomes a plainly bad idea.

If we take a life expectancy of 3 years (already quite optimistic for most gadgets) and a solar insolation of 900 kWh/m² (quite optimistic too, since these things are not lying on a roof), the result is 1,038 gram CO2 per kWh in the worst case scenario (high-efficient mono-crystalline cells produced in the US). That means that it is better for the environment to power a gadget with electricity generated by coal, rather than by a solar panel.

Now what?

All this does not mean that PV solar energy should not be promoted. For one thing, it’s much better using silicon wafers to make energy generating equipment instead of energy guzzling equipment (like computers, mobile phones and car electronics). But some facts have to be faced.

First, solar cells are far from a zero emission technology. Two: solar panels can be a doubtful choice in less sunny regions. Three: solar panels mounted on gadgets are completely insane. Four: solar cells should be recycled. Five: some law or incentive should be introduced to guarantee a life expectancy of 30 years. And if possible, solar thermal power should have priority over solar PV power.

All this does not mean that PV solar energy should not be promoted. But some facts have to be faced.

It should be realized that solar panels first raise the amount of greenhouse gasses before they help lowering them. If the world would embark on a giant deployment of solar energy, the first result would be massive amounts of extra greenhouse gasses, due to the production of the cells.

Use solar panels to produce solar panels

A better strategy would be to use already available solar panels to produce more solar panels. The scientists calculated that the ecological burden of solar panels can be halved if 100 percent of energy in the factories would be delivered by solar energy.

I did not do the calculations for air pollution and heavy metals, but since these are mainly produced by energy use for production, the conclusions must be similar.

© Kris De Decker (edited by Vincent Grosjean)

Leave a Reply

Your email address will not be published. Required fields are marked *