MASU on Machines: Photovoltaic Cell Energy Payback

Welcome back to the MASU on blog and the first thread for 2008

In this series and other unassociated but similarly themed threads there has been considerable discussion on the subject of the energy that is consumed during the production of photovoltaic PV cells compared to the energy they generate over their operational life. There is an encyclopedic volume of what can at best be described as questionable data and the claim that PV cells do not produce as much energy as they consume. So, what is the reality of the situation, what is fact and what is urban myth?

First off let's have a look at the intensity of solar radiation at ground level. The most commonly bandied about figure and one used in many calculations is 1,000 Wm^-2.

So, is that a realistic figure or not?

The chart below shows the amount of solar radiation that reached the ground in Australia on 24^th January 2008.

It is currently summer in Australia and the Sun is above the horizon for pretty close to 14 hours a day while the average energy to reach the ground is around 25 MJm^-2. We can then calculate a value for the average power per square metre as below:

That's considerably less than the commonly used 1,000 Wm^-2 and we will look at the consequences of that later.

Next off we need to look at the efficiency of Photovoltaic PV cells. This varies considerably depending on the quality of the solar cells, the method of construction, their intended use and a host of other parameters. For example those used in satellites and spacecraft are greatly more efficient, reliable, less massive and considerably more expensive than those used on fixed ground based installations. There are also new manufacturing techniques like sliver technology that not only reduced the cost and materials required to manufacture PV cells but also dramatically increases their efficiency over a greatly increased range of conditions.

For the moment let's assume that we are talking of the average cells that are readily available for ground installations and have an efficiency of around 30%. If we combine this with the data from the chart above, gives us an effective power output of 66 Wm^-2.

It's not realistic to think that we could supply the world's total energy needs from PV cells but it is worth calculating the area that would be required to do so in order to give us some sort of idea about the magnitude of the problem we are facing. First off we need to calculate the average power requirement as shown below:

Next we need to calculate the area of PV cells that would be needed to generate the 16 TW currently being consumed:

However, this doesn't take into account the time that the PV cells are in darkness, under cloud cover or other shadows or other parameters that limit the energy the cells can produce. It also doesn't take into account any increase in power demand or efficiency of so far unknown or undeveloped technology.

If we were to utilize PV cells to supply about 10% of the global energy demand and assume they would only be utilized in places where there was close to 12 hours daylight per day there would be a need to manufacture some 50 x 10⁹m² of PV cells. 50 billion square metres of PV cells is one mammoth production run and would definitely require a coordinated global and multi company program. However, considering semiconductors production has been following Moore's law for several decades will this remain an unrealistic target forever?

Getting back to one of the original questions and commonly quoted factoids is the energy returned during the operational life of PV cells really less than the energy used in their production. This is called the Energy Payback of PV Cells and you will find a fairly extensive review of the information available in 2006 if you follow the link. Some of the main points this article covers are:

• Use of Waste Si Wafers: The wafers that are commonly used in the commercial manufacture of PV cells for domestic and terrestrial use are constructed for Silicon wafers that are not suitable for use in the manufacture of semiconductors. Silicon is produced in ingots that are then cut into wafers that are roughly the same size as a CD/DVD. If the wafers have any imperfections, flaws, occlusions etcetera they are not suitable for the production of semiconductors and are therefore treated as waste. However, these wafers are usually suitable for use in the production of PV cells so in reality PV cells are constructed from the waste products of the semiconductor industry. This then raises the question of whether or not the energy consumed in the creation of the wafers should be used in the calculation of energy required to produce PV cells. After all, if these wafers were not used for PV cells they would be just discarded or recycled to produce new Si ingots.
• Centralized Power Generation: Many of the studies into the energy payback time of PV cells assume that they will be used at centralized locations. The current use of PV cells for domestic and industrial power generation utilizes a distributed system where the power they generate is used locally. Surplus power is then distributed to other energy users over the existing power distribution grids.
• Power Plant Construction: Following on from the previous point many calculations include the energy that is utilized in the construction of centralized power stations. However, since we are talking of a distributed system there is no need for the construction of special purpose built installations. With a distributed power generation system much of the infrastructure and building requirements will already be in place and only require minor modifications to have panels of PV cells installed.
• Operation, Monitoring and Maintenance: Again many calculations assumed that you would need to include the ongoing operation, monitoring and maintenance of centralized power generation. However, with a fully automated and distributed system the need for such work is dramatically reduced.

The conclusion that was drawn by The Environmental Engineer paper was that for a distributed domestic power generating system based on PV cells the energy payback time would be somewhere between 2 and 8 years depending on specific installation parameters. When this is compared to the average expected life expectancy of 20 to 25 years a distributed power generating system based on PV cells will produce between 3-12 times the amount of energy that was used in the production, installation and operation.

Sliver Technology & Other New Technologies.

At the time The Environmental Engineer paper was written sliver technology was still in the very early stages of development. The technology is currently at the trial production level with a test manufacturing facility due to come on line sometime in early 2008 with trial installations coming on line a few months later.

What is sliver technology? Basically sliver technology takes an existing 2 mm thick silicon wafer and cuts it into extremely thin slices. These slices are then rotated through 90° and mounted on a backing substrate. This has several positive results:

• Increased Surface Area: By cutting the wafer into slivers that are 50 μm wide and 2 mm thick then rotating them through 90° the effective collecting area that can be created from a single wafer can be increased by some 10-12 fold.
• Electrode Position: With a normal solar cell the connections need to be mounted on the upper and lower surface of the cell. Since sliver technology rotates the slivers by 90° the connections are now made on the sides of each sliver and has the following effects:
• Bidirectional: Because the electrical connections are at the side of each sliver the cells can generate power regardless of which side is exposed to light. This not only makes the mounting easier but can increase the output if a mirror system is employed that illuminates both sides of the cell.
• Output Voltage: With the current generation of mass produced PV cells there is a disproportionate drop in output voltage as the area exposed, intensity and angle of incidence of light varies. Sliver cells can operate over a dramatically increased range of parameters and will generate their designed output voltage with as little 5-10% of their surface exposed to radiation.

Sliver technology has the potential to reduce the energy payback time by anywhere between 5-10 fold. That would bring the energy payback period down something like 3-18 months.

PV cells have been around for something like 120 years but until around 50 years ago remained nothing more than a scientific curiosity. The first PV cell was constructed from selenium with an extremely thin coating of gold that had an overall efficiency of around 1%. This efficiency increased to about 6% in 1954 when Bell Laboratories accidentally found it was possible to construct PV cells from silicon doped with certain trace impurities. Research into PV cell technology has exploded in the last decade or so and efficiency is now approaching the theoretical maximum of 40%. Over the last century there was undoubtedly a time when the energy consumed in the creation of PV cells exceeded the energy they would produce but with current technology this is no longer the case. The energy payback period is currently somewhere between 5-10 years and is likely to drop to somewhere around 3-18 months in the next few years.

Finally I would just like to differentiate between energy payback and financial payback. Energy payback is only concerned with the energy consumed in production compared to the energy returned over the operational life of PV cells. It does not in any way look into the cost of production, monetary reward or offset of the energy produced by PV cells. Origin Energy Australia are currently offering an off the shelf PV cell based power generating system that has a projected financial payback period of around 5-10 years. Admittedly this payback period is reduced by government subsidies, but Origin Energy are close to completing a pilot plant that manufactures sliver technology PV cells. If sliver technology lives up to its potential then we will more than likely see a reduction in the financial payback period of 20-50%. The financial payback period also does not include inflationary pressures on the price of energy. The more expensive energy becomes, the shorter the financial payback period will become.

In conclusion, The concept that the production of PV cells consumed more energy than that they produced during their operational life may well have been true. However, with the increases in efficiency of PV cells and improved production techniques this is no longer the case. Currently the energy payback period is 5-10 years and is likely to drop to as little as 3-18 months in the next few years.

As usual you can read more about this by following the links below.

• Photovoltaic Cells: Wikipedia
• Energy Payback of PV Cells: The Environmental Engineer
• Moore's Law: Wikipedia
• SI Prefixes: Wikipedia
• World Energy Resources & Consumption: Wikipedia