Tuesday, March 2, 2010

Letter to Senate on Lithium Batteries

Lindsay Leveen (as an individual)

Submitted To the Senate Subcommittee For Energy and Water Development

OWT on the Subject of plug-in vehicles that require rechargeable lithium batteries.

An Essay on the Thermodynamics and Economics of Lithium Batteries

My name is Lindsay Leveen. I am a chemical engineer and my interest is to apply my scientific knowledge to alternate energy sources. My graduate work involved the study of thermodynamics. Over the last 35 years my work has been in cryogenics, microelectronic device fabrication, nanotechnology development, fuel cell fabrication, and most recently biotechnology.

Purpose: The purpose of this essay is to provide the subcommittee with reasoning based on thermodynamics why lithium batteries will likely not lower in cost and therefore why plug in passenger vehicles (cars and trucks) will probably not make any significant dent in the consumption of gasoline and diesel. I wish to prevent the waste of precious resources on a technology that I believe is headed toward a dead end.

I have no commercial interest in any energy or battery technology and am writing this essay as a concerned citizen to inform the Senate Subcommittee on Energy and Water Development of the severe thermodynamic limitations of Lithium Secondary Batteries and of the probable long term unaffordable economics associated with plug-in passenger vehicles that will rely on them. Much of this report is taken from my presentations, reports, publications and blogs www.greenexplored.com I have produced in recent years.

Thermodynamics – definition: “the science concerned with the relations between heat and mechanical energy or work, and the conversion of one into the other: modern thermodynamics deals with the properties of systems for the description of which temperature is a necessary coordinate.” (dictionary.com).

Moore’s Law and Learning Rates for Technologies: Gordon Moore one of the founders of Intel Corporation, postulated that semiconductor integrated circuits would enjoy a doubling in performance in a period of every 18 months. This rate of learning allows performance to be improved exponentially with time for the same original cost.

Many technologies that engineers and scientists develop need a “Moore’s Law” in order to improve their performance and correspondingly their economics to capture vast markets. Most efforts around the improvement of alternate energy technologies vis a vis competing with fossil fuels have not yielded these “Moore’s Law” rates of learning. In particular for the past decade as much as six billion dollars has been spent without any real success toward the “learning curve” of PEM fuel cells. Much of these six billion dollars was appropriated by the Federal Government. The learning curve for PEM fuel cells over the past decade yielded a yearly learning rate of less than 2%. By comparison the Moore’s Law yearly learning rate for integrated circuits has averaged over 40% for more than three decades.

My experience with Moore’s Law: For almost twenty years I directed teams of engineers that designed state of the art Integrated Circuit (IC) fabrication facilities that helped drive this rapid rate of learning and therefore cost improvement in computers and other electronic devices. A simple explanation for the high learning rates in IC fabrication is that the technology was neither constrained by thermodynamics nor reaction kinetics but simply by the line width of the circuits within the ICs. To drive Moore’s law in IC fabrication improvements in lithography, higher purity gases for deposition, implantation, and etch, as well as the occasional increase in the size of wafer being fabricated were needed.

Moore’s Law, Thermodynamics and Lithium Batteries: To drive the learning rate in PEM fuel cells and similarly lithium secondary batteries both thermodynamic and reaction kinetic constraints have to be overcome. The reason why thermodynamics places such constraints is that the functioning of these systems depends on chemical reactions. Thermodynamics determines how much useful energy can be derived from a chemical reaction. But we know that the thermodynamic constraints cannot be overcome as the laws of thermodynamics are inviolable. ICs do not undergo chemical reactions to function, but all batteries and fuel cells do involve chemical reactions to deliver energy. It is these chemical reactions that are limiting the possible learning rate.

The Resulting Economic Problem: Significant effort and much money is now being spent on advanced batteries for plug-in full electric or plug-in hybrid vehicles. Such vehicles will require between 10 kilowatt hours and 50 kilowatt hours of stored electricity if the range of the vehicle purely propelled on stored electricity is to be between 40 and 200 miles. Lithium chemistry based secondary (chargeable) batteries presently offer the best performance on a weight and volume basis and therefore represent the best “hope” for a “Moore’s law” to solve the world’s addiction to fossil oil. Sadly “hope” is not a winning strategy. Present costs of such battery packs at the retail level range from $800 per kilowatt hour of storage to over $2,000 per kilowatt hour of storage. One can purchase a 48 volt 20 amp hour Ping Battery for an electric bicycle directly from this Chinese “manufacturer” for less than $800 delivered by UPS to any address in the USA. A123 offers a battery system that will modify a standard Prius to a 5 kilowatt hour plug-in Prius for $11,000 or around $2,200 per kilowatt hour fully installed by a service station in San Francisco. The Ping battery delivers much less instantaneous power (watts) and that is the reason their batteries are less expensive on a stored energy basis (watt hours) than are the A 123 batteries. Both the Ping and the A123 batteries claim safety and claim to be manufactured with phosphate technology that will neither short circuit nor burn.

Economic Case Study The Example The Standard Prius vs Plug-in Prius: The following is an economic analysis of a standard Prius versus a plug-in Prius using A 123’s lithium battery pack;

The standard Prius will get 50 MPG and let’s assume that the driver drives 12,000 miles a year. The standard Prius driver will need to purchase 240 gallons a year of gasoline at an estimated cost of $720 per year with gasoline at selling for $3 per gallon. If the driver purchased the A 123 plug-in system and can recharge the system at home and at work such that half the mileage driven in a year is on batteries and half is on gasoline the driver will save $360 a year on gasoline. The driver will need to buy some 2,000 kilowatt hours a year of electricity from the grid in order to save this gasoline. At 10 cents per kilowatt hour the driver will spend $200 a year for electric power and will therefore only enjoy $160 a year in net operating savings. The $11,000 set of batteries have a maximum expected life of 8 years and the owner must set aside $1,375 a year for battery replacement without accounting for the time value of money. The battery replacement cost is simply too expensive to justify the savings in gasoline. How high do gasoline costs have to rise and how little do batteries have to cost to make the plug in viable? Let’s assume gas prices reach $6 per gallon and electricity remains at 10 cents a kilowatt hours we have a yearly operating savings of $520. These savings will still be far short of the money needed for battery replacement.

The A 123 batteries will need to drop to 15% of their present cost to make the proposition of converting a Prius to a plug -n “worthwhile”. To reach this cost target in a decade one needs a yearly learning rate of approximately 26%. With 35 years of work experience, I have concluded that in the best case of battery costs (no inflation in raw materials) a 4 or 5% yearly learning rate could be achieved over the next decade. But if we believe that gasoline will double then we also have to assume that plastics, copper, cobalt, nickel, graphite, etc. will also double in unit cost. As raw materials account for three quarters of the manufacturing cost of lithium batteries the inflation adjusted cost will increase at a higher yearly rate than the learning rate will lower costs. My prediction is therefore that lithium secondary batteries will likely cost more per unit of energy stored in 2020 than they do today.

Toyota is a company well known for its cars with improved fuel economy and therefore is a master of thermodynamics and must have “optimized” the cost and performance of its batteries in the standard Prius deploying a relatively small battery pack and with the choice of Nickel Metal Hydride chemistry rather than lithium chemistry. While Toyota may be experiencing safety problems no one can fault this company on fuel efficiency. Other car companies such as Ford have also chosen Nickel Metal Hydride as their hybrid car battery platform. Fisker and GM are touting plug in hybrids with lithium batteries and are much more aggressive in their claims of cost improvement and their ability to drive “Moore’s Law” in their battery systems. My educated guess on all of this is that Toyota, Ford and the car manufacturers that stick with smaller nickel metal hydride battery systems and the traditional non plug-in hybrid will sell tens of millions of such vehicles over the next decade. Renault, GM, Fisker, Tesla, and others who go for plug-in hybrids or full electric vehicles will only sell a few tens of thousands of vehicles in the next decade. I simply believe we will not have “Moore’s Law” at play here but have a very fractional Moore’s Law that holds.

Argonne National Labs published an exhaustive review of the materials and associated costs of lithium batteries back in May of 2000. http://www.transportation.anl.gov/pdfs/TA/149.pdf The total material cost for the cell was estimated at $1.28 and the total manufacturing cost of the cell including overhead and labor was estimated at $1.70. This Argonne report is perhaps the best report written on the economics associated with lithium battery fabrication. Actually had folks read this report back in 2000 they would have realized that the learning curve for lithium batteries would be painfully slow. Materials just make up far too much of the battery cost and the quantity of materials is fixed by the chemistry. Therefore economies of scale could not drive a Moore’s Law type rate of learning and a very fractional Moore’s Law resulted. In the early years of lithium cell development from approximately 1990 to 2000, the improvements in chemistry and in economies of scale did allow the technology to enjoy a Moore’s Law type learning rate and it has been reported that costs of an 18650 cell reduced from $18 to $2 per cell in that decade. Unfortunately the technology has now hit an asymptote in their cost reduction curve.

By doing a Google search on an 18650 lithium ion battery I came across this link http://www.batteryjunction.com/li18322mahre.html . This site lists a selling price of $5.29 each for 200 or more cells. The cells are 3.7 volts with 2.2 amp hours so they are capable of holding 8.1 watt hours of energy from full charge to discharge. Expressed in cost per kilowatt hour of nominal capacity these loose cells cost around $650. My guess is that if applied today’s costs of cobalt, nickel, lithium, lithium salts, plastics, copper, graphite, and other constituent materials that make up a cell, the material cost in November 2009 compared with May 2000 have increased by more than 150% and a current estimate of the materials used in the Argonne labs report will show cost of about $3 per cell versus $1.28 back in May 2000. Hence this company sells the cells for $5.29 each. From my previous analysis of the probable learning rate I would not surprised if in 2020 the selling price per 18650 lithium cell is as high as $6 rather than as low as $3.

Conclusion: Lithium batteries are and will remain best suited for items as small as a cell phone and as large as a bicycle. The cost relative to performance or these batteries will likely not improve by much in the coming decade. Although some standard hybrid vehicles may use lithium batteries with low capacity, their cost will remain high. Also plug-in vehicles that have a range longer than 10 miles using battery power will likely not penetrate the market significantly. Given the likely scenario that plug-in passenger cars and trucks based on lithium battery technology will not reduce US consumption of gasoline and diesel fuel in large measure, I am asking the subcommittee to limit the funds that the US government will appropriate for research and development of this technology.

Thank you

Lindsay Leveen