The Economist has a post on research from the Utrecht University into the energy investment required for solar panels and the amount of carbon emissions involved - How clean is solar power?. You can see how far solar PV has come in terms of energy return on investment when the numbers being quoted for new panels are around 15 now (taking less than 2 years to return the initial energy investment).

Wilfried van Sark, of Utrecht University in the Netherlands, and his colleagues have ... calculated the energy required to make all of the solar panels installed around the world between 1975 and 2015, and the carbon-dioxide emissions associated with producing that energy. They also looked at the energy these panels have produced since their installation and the corresponding amount of carbon dioxide they have prevented from being spewed into the atmosphere. Others have done life-cycle assessments for solar power in the past. None, though, has accounted for the fact that the process of making the panels has become more efficient over the course of time. Dr Van Sark’s study factors this in. [The team] found that solar panels made today are responsible, on average, for around 20 grams of carbon dioxide per kilowatt-hour of energy they produce over their lifetime (estimated as 30 years, regardless of when a panel was manufactured). That is down from 400-500 grams in 1975. Likewise, the amount of time needed for a solar panel to produce as much energy as was involved in its creation has fallen from about 20 years to two years or less. As more panels are made, the manufacturing process becomes more efficient. The team found that for every doubling of the world’s solar capacity, the energy required to make a panel fell by around 12% and associated carbon-dioxide emissions by 17-24%.

PhysOrg has an article linking recent recessions with declines in the EROI of major fuel sources - Declining energy quality could be root cause of current recession.

Many economists have pointed to a bursting real estate bubble as the initial trigger for the current recession, which in turn caused global investments in U.S. real estate to turn sour and drag down the global economy. King suggests the real estate bubble burst because individuals were forced to pay a higher and higher percentage of their income for energy—including electricity, gasoline and heating oil—leaving less money for their home mortgages.

In economic terms, the quality of the nation's energy supply is referred to as Energy Return on Energy Investment (EROI). For example, if an oil company uses a 10th of a barrel of oil to drill, pump, transport and refine one barrel of oil, the EROI for the refined fuel is 10.

"Many economists don't think of energy as being a limiting factor to economic growth," says King, a research associate in the university's Center for International Energy and Environmental Policy. "They think continual improvements in technology and efficiency have completely decoupled the two factors. My research is part of a growing body of evidence that says that's just not true. Energy still plays a big role."

In a paper published this November in the journal Environmental Research Letters, King introduced a new way to measure energy quality, the Energy Intensity Ratio (EIR), that is easier to calculate, highly correlated to EROI and in some ways more powerful than EROI. EIR measures how much profit is obtained by energy consumers relative to energy producers. The higher the EIR, the more economic value consumers (including businesses, governments and people) get from their energy.

When King plots EIR for various fuels every year since World War II, the graphs indicate two large declines, one before the recessions of the mid-1970s and early 1980s and the other during the 2000s, leading up to the current economic recession. There have been other recessions in the U.S. since World War II, but the longest and deepest were preceded by sustained declines in EIR for all fossil fuels.

EIR is proportional to EROI, meaning they rise and fall together, but the basic data behind the EIR calculations come out annually as opposed to every five years for EROI. EIR also gives insight into different parts of the supply chain such as at the refinery or at the gas pump, which are harder to study with EROI.

King's analysis suggests if EIR falls below a certain threshold, the economy stops growing. For example, in 1972, EIR for gasoline was 5.9 and in 2008 it was 5.5. During times of robust economic growth, such as the 1990s, EIR for gasoline was well over eight. Compare that to some estimates of EROI and EIR for corn ethanol of around one, and it's clear why corn ethanol has been widely criticized as a low quality energy source.

To get the U.S. economy growing again, King says Americans will have to produce and use energy more efficiently. That's essentially what the U.S. did after the last energy crisis by raising fuel efficiency standards for cars, increasing use of natural gas for electric power generation and developing new technologies such as Enhanced Oil Recovery to coax more oil out of the ground.

The Oil Drum has a post from Tom Konrad looking at the concepts of EROI and EIRR and how they impact the transition to a post-oil world - Managing the Peak Fossil Fuel Transition: EROI and EIRR.

Energy keeps our economy running. Energy is also what we use to obtain more energy. The more energy we use to obtain more energy, the less we have for the rest of the economy.

The concept of Energy Return on Investment (EROI), alternatively called Energy Return on Energy Invested (EROEI) has been widely used to quantify this concept. The following chart, from a SciAm paper, shows the EROI of various sources of energy, with the tan section of the bar representing the range of EROIs depending on the source and the technology used. I've seen many other estimates of EROI, and this one seems to be on the optimistic (high EROI) end for most renewable energy sources.

The general trend is clear: the energy of the future will have lower EROI than the energy of the past. Low carbon fuels such as natural gas, nuclear, photovoltaics, wind, and biofuels have low EROI compared to high-carbon fuels such as coal and (formerly) oil.

The graph also clearly shows the decline in the EROI over time for oil. Other fossil fuels, such as coal and natural gas, also will have declining EROI over time. This happens because we always exploit the easiest resources first. The biggest coal deposits that are nearest to the surface and nearest to customers will be the first ones we mine. When those are depleted, we move on to the less easy to exploit deposits. The decline will not be linear, and new technology can also bring temporary improvements in EROI, but new technology cannot change the fact that we've already exploited all the easiest to get deposits, and new sources and technologies for extracting fossil fuels often fail to live up to the hype.

While there is room for improvement in renewable energy technologies, the fact remains that fossil fuels allow us to exploit the energy of millions of years of stored sunlight at once. All renewable energy (solar, wind, biomass, geothermal) involves extracting a current energy flux (sunlight, wind, plant growth, or heat from the earth) as it arrives. In essence, fossil fuels are all biofuels, but biofuels from plants that grew and harvested sunlight over millions of years. I don't think that technological improvements can make up for the inherent EROI advantage of the many-millions-to-one time compression conveys to fossil fuels.

Hence, going forward, we are going to have to power our society with a combination of renewable energy and fossil fuels that have EROI no better than the approximately 30:1 potentially available from firewood and wind. Since neither of these two fuels can come close to powering our entire society (firewood because of limited supply, and wind because of its inherent variability.) Also, storable fuels such as natural gas, oil, and biofuels all have either declining EROI below 20 or extremely low EROI to begin with (biofuels). Energy storage is needed to match electricity supply with variable demand, and to power transportation.