Energy Payback from Photovoltaics: Problems in Calculation
by Jeff Vail
Does solar energy—specifically photovoltaic (PV) panels—ever produce as much energy as the energy that was initially invested in their manufacture? Industry, academia, and government all seem to be in agreement that the answer is “yes.” (1)(2)(3) The consensus seems to be that PV produces as much energy as was used in its creation in a time period of 1-5 years, allowing PV to produce between 6 and 30 times more energy over its life than was used in its creation. These two answers—that PV produces more energy than is used in manufacture, and that PV provides an Energy Return on Energy Invested (EROEI) of between 6:1 (2) and 30:1 (2)—suggest that photovoltaics can be and should be a cornerstone of our efforts to replace our reliance on non-renewable fossil fuels.
There are serious problems, however, with the methodology used at present to calculate the EROEI of solar panels. Some authors claim that life-span EROEI for photovoltaics is as high as 50, but provide no information for how that figure is calculated. (4) Others, such as Clarion University’s calculations, take a very limited view of energy invested in PV production, accounting only for energy use of the manufacturing plant itself. Under these assumptions, they understandably arrive at a very optimistic EROEI of 6:1 to 31:1. (1) So what energy inputs are not being accounted for in such a calculation? Let’s work backwards:
- Installation: PV does not good sitting in the factory. It must be installed, and this takes labor. There are various ways of accounting for the energy represented by such labor, but it certainly takes energy.
- Transportation: PV has to get to the installation site. Efficient manufacture is only possible if it is centralized, but this means that it must be shipped—usually by truck, which requires both the fuel directly consumed by shipping, plus the energy consumed in the entire chain of operation necessary to construct the truck, as well as the labor cost of the driver, which also represents an energy input.
- Manufacturing plant: EROEI calculations usually account for the energy consumption of the manufacturing plant, but not for the construction of the manufacturing plant itself, as well as the construction of all the machines used on the PV assembly line (PV advocates often point out that silicon is the most abundant element on earth and therefore requires very little energy to acquire—but this is NOT true for the highly advanced manufacturing machinery necessary to create PV cells, usually made from metals that require great energy input for extraction). If we take the total energy required to create one PV manufacturing plant as well as its expected lifetime production, we can then calculate how much of that energy should be attributed to a given quantity of PV panel.
- Labor: One of the key components in the production of PV panels is human input, and yet this energy cost is not accounted for in standard EROEI calculations. I’m not referring to the actual calories expended operating an assembly line, or answer the phones in the front office, but rather the energy consumed in the course of these people’s daily lives—energy that must be accounted for because it is part of the support structure necessary to create a PV panel. No employees, no PV.
These embodied energy costs in the creation of a PV panel (called “emergy”) are difficult to calculate. We can regress infinitely, eventually going so far as to account for the portion of energy consumed by a rice farmer in China in order to fill the belly of a Merchant Marine captain shipping machine parts across the Pacific, ad infinitum. How do we actually get a composite sense of the total embodied energy in PV production? One way—and certainly not a perfect way—is to use the market’s ability to set prices as an equivalent for embodied energy. This is what I am calling “Price-Estimated EROEI Theory.” It basically suggests that the most accurate representation of the total energy embodied in ANY product is the price of that product. In our example above, the energy required to install PV can be accounted for by the cost of that service. The energy required to transport, to build a manufacturing plant, to employ workers, etc.—all component energy contributions in the production of PV increase the market price of the resulting product.
So what is the Price-estimated EROEI of PV? If we accept that the price of an installed PV system is representative of the energy used, then we can compare that price with the quantity of energy produced over the lifetime of that system (which also has a market price) and reach an EROEI ratio. There are variables involved here, but when we use market-price to account for the full spectrum of energy “invested” in PV, we reach an EROEI of approximately 1:1 (*see full calculations below). This is dramatically different than the 6:1, 30:1, or 40:1 suggested by most sources. Which figure should we rely upon? While I recognize that price-estimated EROEI is not a perfect calculation, at least it attempts to account for the full spectrum of energy inputs, and the precautionary principle suggests that we should err on the side of this number (1:1) as opposed to the quite optimistic figures coming from the PV industry or the government.
Ultimately there is only one way to definitively answer this questions: The bootstrap challenge. I have previously stated that when I see an ethanol plant that distills their ethanol USING ethanol (not natural gas or coal), then I will seriously reconsider the merits of that alternative energy source. Likewise, when I see a PV production plant that is powered entirely by PV, containing machines manufactured at plants powered entirely by PV, machines composed of materials mined, refined, and shipped entirely under PV power, etc., then I will believe that PV has an EROEI greater than 1:1. With an EROEI like 30:1, this should be no problem . . . so the fact that this is not the case is yet another argument, at least in my mind, that reality stands closer to the 1:1 figure.
EROEI is not just a nifty academic exercise. The outcome of the debate on EROEI—whether for PV panels, ethanol production, nuclear fission—is critically important for the future of our economy and society. Regardless of the exact timeline, it is not seriously disputed that non-renewable energy sources such as oil, gas, and coal—all with high EROEI—are running out. There is a commonplace assumption that we will create alternatives to replace them, but at present these alternatives—from PV to ethanol—are all being produced with the very fossil fuels that are disappearing. When they are effectively gone, only energy sources with an EROEI of greater than 1:1 will be viable—and even then, our economy, with its demand for constant growth, cannot survive on energy with an EROEI of 2:1 or 5:1. For that reason, it is critical that we more carefully address this EROEI debate today. If alternative, truly renewable sources of energy cannot match—and eventually improve upon—the EROEI of today’s energy sources, then we must conduct a serious reappraisal of the fundamental structure of our society. My analysis suggests that we must do exactly that.
* CALCULATION: 2 KW complete PV system installed in Phoenix quoted at $16,000 (before any tax rebates or incentives, grid-intertie only,NOT including battery storage)(5). In Phoenix (optimal location), this generates 4,000 KW-hours of electricity per year (5). At prevailing Phoenix rate of electricity ($.10/KW-hour) this is $400/electricity per year. This produces a pay-back time of 40 years if we do not account for the time-value of money. For the purposes of this calculation I will be very conservative and find that actual inflation will equal TVM over this 40-year period. The quoted PV panels have a life-expectancy of 40-years (again, this is conservative as this “complete” system ignores battery storage, which would dramatically decrease the aggregate life expectancy). The resulting price-estimated EROEI of PV solar is 1:1.