Report: Some 70 companies trying to partner with Linc Energy on developing massive Australian shale oil prospects
Eaton and Cummins introduce powertrain package for North American heavy-duty trucks offering 3-6% better fuel economy

Stanford study quantifies energetic costs of grid-scale energy storage over time; current batteries the worst performers; the need to improve cycle life by 3-10x

A plot of ESOI for 7 potential grid-scale energy storage technologies. Credit: Barnhart and Benson, 2013. Click to enlarge.

A new study by Charles J. Barnhart and Sally M. Benson from Stanford University and Stanford’s Global Climate and Energy Project (GCEP) has quantified the energetic costs of 7 different grid-scale energy storage technologies over time. Using a new metric—“Energy Stored on Invested, ESOI”—they concluded that batteries were the worst performers, while compressed air energy storage (CAES) performed the best, followed by pumped hydro storage (PHS). Their results are published in the RSC journal Energy & Environmental Science.

As the percentage of electricity supply from wind and solar increases, grid operators will need to employ strategies and technologies, including energy storage, to balance supply with demand given the intermittency of the renewable supply. The Stanford study considered a future US grid where up to 80% of the electricity comes from renewables.

Only about 3% currently is generated from wind, solar, hydroelectric and other renewable sources, with most of the electricity produced in the United States currently coming from coal- and natural gas-fired power plants, followed by nuclear according to data from the US Energy Information Administration (EIA) .

They quantified energy and material resource requirements for currently available energy storage technologies: lithium ion (Li-ion), sodium sulfur (NaS) and lead-acid (PbA) batteries; vanadium redox (VRB) and zinc-bromine (ZnBr) flow batteries; and geologic pumped hydroelectric storage (PHS) and compressed air energy storage (CAES).

The current total energy storage capacity of the US grid is less than 1%, according to Barnhart. What little capacity there is comes from pumped hydroelectric storage, which works by pumping water to a reservoir behind a dam when electricity demand is low. When demand is high, the water is released through turbines that generate electricity.

By introducing new concepts, including energy stored on invested (ESOI), we map research avenues that could expedite the development and deployment of grid-scale energy storage. ESOI incorporates several storage attributes instead of isolated properties, like efficiency or energy density.

Calculations indicate that electrochemical storage technologies will impinge on global energy supplies for scale up—PHS and CAES are less energy intensive by 100 fold. Using ESOI we show that an increase in electrochemical storage cycle life by tenfold would greatly relax energetic constraints for grid-storage and improve cost competitiveness. We find that annual material resource production places tight limits on Li-ion, VRB and PHS development and loose limits on NaS and CAES.

This analysis indicates that energy storage could provide some grid flexibility but its build up will require decades. Reducing financial cost is not sufficient for creating a scalable energy storage infrastructure. Most importantly, for grid integrated storage, cycle life must be improved to improve the scalability of battery technologies. As a result of the constraints on energy storage described here, increasing grid flexibility as the penetration of renewable power generation increases will require employing several additional techniques including demand-side management, flexible generation from base-load facilities and natural gas firming.

—Barnhart and Benson

The first step in the study was to calculate the cradle-to-gate embodied energy—the total amount of energy required to build and deliver the technology—from the extraction of raw materials, such as lithium and lead, to the manufacture and installation of the finished device.

To determine the amount of energy required to build each of the five battery technologies, the authors used data collected by Argonne National Laboratory and other sources. The data revealed that all five battery technologies have high embodied-energy costs compared with pumped hydroelectric storage.

After determining the embodied energy required to build each storage technology, the next step was to calculate the energetic cost of maintaining the technology over a 30-year timescale. To quantify the long-term energetic costs, Barnhart and Benson came up with a new mathematical formula they dubbed ESOI, or energy stored on investment.

ESOI is the amount of energy that can be stored by a technology, divided by the amount of energy required to build that technology. The higher the ESOI value, the better the storage technology is energetically.

—Charles Barnhart

The results showed that CAES had the highest value: 240. In other words, CAES can store 240 times more energy over its lifetime than the amount of energy that was required to build it. (CAES works by pumping air at very high pressure into a massive cavern or aquifer, then releasing the compressed air through a turbine to generate electricity on demand.) PHS followed at 210.

The five battery technologies fared much worse. Lithium-ion batteries were the best performers, with an ESOI value of 10. Lead-acid batteries had an ESOI value of 2, the lowest in the study.

The best way to reduce a battery’s long-term energetic costs would be to improve its cycle life, the Barnhart said. Pumped hydro storage can achieve more than 25,000 cycles; none of the conventional battery technologies featured in the study has reached that level. Lithium-ion is the best at 6,000 cycles, while lead-acid technology is at the bottom, achieving a mere 700 cycles.

The most effective way a storage technology can become less energy-intensive over time is to increase its cycle life.Most battery research today focuses on improving the storage or power capacity. These qualities are very important for electric vehicles and portable electronics, but not for storing energy on the grid. Based on our ESOI calculations, grid-scale battery research should focus on extending cycle life by a factor of 3 to 10.

—Sally Benson

In addition to energetic costs, Barnhart and Benson also calculated the material costs of building these grid-scale storage technologies. In general, they found that the material constraints aren’t as limiting as the energetic constraints. However, PHS has a different type of challenge—the number of geologic locations conducive to pumped hydro is dwindling, and those that remain have environmental sensitivities, Barnhart noted.

A primary goal of the study was to encourage the development of practical technologies that lower greenhouse emissions and curb global warming, Barnhart said. Coal- and natural gas-fired power plants are responsible for at least a third of those emissions.

I would like our study to be a call to arms for increasing the cycle life of electrical energy storage. It’s really a basic conservative principal: The longer something lasts, the less energy you’re going to use.

—Charles Barnhart

The study was supported by GCEP and its sponsors: ExxonMobil, GE, Schlumberger and DuPont.


  • Charles J. Barnhart and Sally M. Benson (2013) On the importance of reducing the energetic and material demands of electrical energy storage. Energy Environ. Sci., doi: 10.1039/C3EE24040A


Kit P

“Only about 3% currently is generated from wind, solar, hydroelectric and other renewable sources, ”

That is not right, let me check:

For 2011:

Solar = 0.05%
Wood + biomass = 1.5%
Geothermal = 0.4%
Hydroelectric = 8%
Wind 2.9%

About 13% is renewable energy. The is not the first useless junk science out out Stanford's Environmental Engineering department.

Hydroelectric stores energy behind the dam and wood stores energy in a big pile of wood. This leaves wind to worry about. Stanford has very good wind resources but we never see a wind farm near this elite university.

The first problem with wind is that the elite who think it is a good idea are against it where they live.

The second problem with wind that most of the US population and power demand corresponds to places with poor wind resources. You can not store what you do not have. This leaves places like Iowa who could care less about research out of loony bins like Stanford

“A primary goal of the study was to encourage the development of practical technologies that lower greenhouse emissions and curb global warming ”

How stupid are these people? Let's say that a 100 MWe wind farm produces at full power for 7 hours a day for 700 MWh. The result is 700 MWh not made with fossil fuel. If for two hours a da, demand was such that only 90 MWh was being used. You could spend a lot of money to store the energy, or you could spend the money on building a wind farm some place else or renewable energy that comes with storage.

Nick Lyons

We need low cost, walk-away safe, distributed nuclear base load. Molten salt reactors are the energy future. The nation that figures this out first (China?) is going to get a huge advantage.

Smart grids with integrated storage and long-range transmission of renewable power is a very expensive way to go. Developing economies will build coal plants instead. We need a low-cost nuclear power option.



"..who could care less about research out of loony bins like Stanford"

Certainty not the Silicon Valley Stanford was/is key in creating.



I agree LFTRs are a good idea, China and India are working with thorium, Norway and Russia are too. We had programs until Congress pulled the funding many years ago.


"Energy Stored on Invested, ESOI” & "cradle-to-gate embodied energy"

They're talking about using batteries newly made for the sole purpose of grid storage, but what about using old batteries like those recycled from BEVs? These would still have cycle life left in them.


Also, what about using heat as a storage medium?


PHS may have a wider geographical range than first thought.


Why do we even pay attention to Exxon funded research? It's got a "Big Oil Approved" stamp on it.


Indeed the Molten Salt Reactor is one of our last hopes of reducing carbon emission for base electrical loads. Unfortunate I see that the earliest it can come on-line is around year 2025 in China with a small number of trial deployments.

If the Chinese program get killed by political and industrial intrigues then we will have nothing. I would feel better if at least three countries are independently researching this technology. Right now only the Chinese program appears serious.


In terms of electrical energy required per nominal kWh of battery storage, the ESOI of li-ion batteries would be about 40 for the life time electrical energy out to electrical energy in. The life time electrical energy out to the total energy in is about 12 as the article states but only if all the energy comes from fuel - 75% to generate electricity and 25% for fuel for mining operations.

This distinction should have been made in the article.

I would be more interested in an article on how much energy storage, electrical, thermal, etc., will we need.

Kit P

"This distinction should have been made in the article."


You may want to back and read the CM LCA. It does consider ghg associated with all stages of production including mining.

This LCA is one of the best I have seen. Of course models and reality are two different things. If you power comes form a coal or NG plant, then BEV are not not a better environmental choice. If the power comes from a low carbon source, then it would be a better choice.


ai vin provided many excellent additional reference non-battery storage options.

Two US Molten Salt Reactors were run and NEVER expanded, but left to die - which may say enough.

There are plenty of ways to provide light water emergency reactor cooling water pressure remotely, but arrogant nuclear people have failed to do this for fifty years and are out of second chances.

Kit P


Commercial reactors being built by Russia, China, and India are LWR. You might ask yourself why. The real world is so much more interesting than the make believe world of college professors.

“We need a low-cost nuclear power option.”

We already have it. LWR are generally cheaper than coal-based depending on the location.


@KitP; I was referring to the CMU life cycle analysis only for the energy, ~ 470 kWh per kWh, required to manufacture a Li-ion battery and the source of this energy.

Yes, the LCA for GHGs is also quite good and it could bring some sanity to GHG discussions.


If decommissioning/radioactive waste correcting costs are included - nuclear becomes the most expensive electric power, even before the trademark cost overruns.

Kit P


Thank you for the clarification. The three most important attributes of a LCA are location, location, location. BEV may be a good idea in France or South Korea but not where I live.

For renewable energy and nukes extending the life of a project reduces both the environmental impact and the cost per kwh.

“If decommissioning/radioactive waste correcting costs are included - nuclear becomes the most expensive electric power, ”

In the US those cost have always been included. For the last ten years the cost of making power with nukes is less than with coal. I did mention location. It is unlikely that a new nuke could beat a new coal plant sitting on strip mine in the Powder River Basin. However, when you ship the coal from Wyoming to Georgia, new nukes are cheaper.

China started building nukes when they could no longer produce enough coal with slave labor coal. A new nuke plant is cheaper in China when they are buying the coal from West Virginia.

“There are plenty of ways to provide light water emergency reactor cooling water pressure remotely, but arrogant nuclear people have failed to do this for fifty years and are out of second chances. ”

US LWRs have not failed to provide cooling during a normal shutdown or during an abnormal occurrence such a loss of offsite power. At TMI, the safety systems worked normally but operators turned them off allowing a gas bubble to displace cooling water.

“walk-away safe ”

Safety is about not hurting people. Core damage is nothing more than equipment damage. The key is to provide enough time for people to walk away. If an earthquake occurs, you want the school your children are in to not fall down around them. If the building must be torn down afterward, it is just money.

Emergency Core Cooling Systems (ECCS - see 10CFR50 for definitions) for LWR are a combination of active and passive systems. With an extended loss of AC power, the earliest we could expect to see core damage is 4 hours. The reactor with a passive system in Japan failed first. One system that uses a steam turbine driven pump ran for 53 hours days after it was designed to run. After the core is damaged, then the fission products have to escape the containment. Again this takes time.

The result is that there is plenty of time to walk away. The interesting thing about irrational fear is that if you tell people radiation is coming they will evacuate. Tell them a wall of water of unpredictable size is coming and they get out the video camera to document the moments before their deaths.


Kit P insults people, it is his way of trying to obtain the illusion of dominance. The stories on here are good, but his comments make it not worth while.


Game changer: Which energy/transport technology leads to least waste? Answer: Plug-in Hybrid vehicles by which households gain the choice to use stored electricity for driving or for household uses; the means to more closely monitor all energy consumption; an economic incentive to drive less, walk, bike and use mass transit more and support local economic development. A modest PHEV battery pack is a better match with a likewise modest rooftop photovoltiac solar array, a lifesaving portable power supply in an emergency, and the advantage of using any available fuel to recharge the battery in a grid failure.
Bottom Line: Plug-in Hybrid technology offers the most benefits and advantages to curb GHG.


Molten salt solar eliminates all these issues, by providing load balancing from solar to balance wind. It's incredibly logical for the US west and south. No nukes needed, no storage needed.


Why no consideration of flywheel systems? It's already in use as grid storage.

"Stanford has very good wind resources but we never see a wind farm near this elite university."

I grew up on Stanford campus. I don't recall it being a very windy place, except for winter storms.


Concentrated solar with molten salt storage, flywheels - they are all valid points. The authors did not mention them because they loved to knock down straw-man opponents.


As Kit mentions in his first post biomass is one way to store renewable energy so I asked myself 'what's the potential?'

Let's take one thing that is produced yearly, like straw because I've already done some research into straw bale houses, and see what we get. About 140 million tons (128 million tonnes) of straw are produced each year in North America (from 1980's & 90's statistics for wheat, rice, oats, rye and barley crops). In many areas straw is still burned in fields, producing significant air pollution. [In California more than 1 million tons of rice straw were burned each fall in the early 1990s, generating an estimated 56,000 tons (51,000 tonnes) of carbon monoxide annually — twice that produced from all of the state’s power plants!] Straw can be compressed for transport: Straw pellets have a bulk density of >600kg/m3 & straw briquettes are ~1000kg/m3 and it has an energy density of 4.4kwh/kg.

128,000,000,000kg X 4.4kwh = 563,200,000,000kwh of stored energy each year just from straw.


Axion Power claims 2500 cycles to failure with 100% discharge. Firefly Energy has 10000 cycles to failure with 30% discharge. Since weight and volume generally are not a huge issue with grid storage, put in massive quantities of Firefly's battery.

Kit P

“I grew up on Stanford campus. ”

Check a reference on wind resources. Of course the SF Bay area is not known for hang gliding, windsurfing, or sailing.

“In California more than 1 million tons of rice straw were burned each fall in the early 1990s ”

Here is a link for those who like pictures.

“Wadham 30 MW
Operation: 1989
Fuel: rice hulls, rice straw
This is the world's largest power plant of its kind and burns up to 600 tpd of rice hulls and rice straw.”

Many waste wood/biomass power plants were built to solve a environmental issues.


A lot has been done, but there is still more that can be done.

The comments to this entry are closed.