Nixon and Dr. Schlachter on the “No Moore’s Law for Batteries” debate.
Note from the Publisher: Due to interest from our audience, this article, which was originally published on May 22, 2015 has been republished.
In my last column published earlier this week titled Electric Dreams’ and Moore’s Law for Batteries, I would like to start by thanking Dr. Fred Schlachter for making time to comment in detail on my previous column with the following:
Thanks for mentioning my opinion piece from April 2013 Proceedings of the National Academy of Sciences, the title of which is “No Moore’s Law for Batteries.” The Editors removed the quotes I had placed around “Moore’s Law'” This “law” is an empirical finding which has become a self-fulfilling prophecy, rather than a law of nature, such as the Second Law of Thermodynamics.
Let me remind the reader that “Moore’s Law” behavior is exponential, evidenced by a straight line on a semi-log plot, or, in popular vernacular, doubling over many time periods with the same time period for each doubling (as, for example, doubling every two years for the past forty years). Simply doubling over some time period is NOT exponential, it is simply doubling. Doubling over several time periods (the same time period in each period) is required to demonstrate exponential behavior. At least four or five time periods are necessary, not just one or two.
As I made clear in my article, the exponential increase of computer power over a long period of time is due to advances in lithography and technology, and has no fundamental meaning, nor is an exponential increase to be expected in any other area of technology. In fact, ongoing improvements in processors will slow and eventually stop as feature sizes approach atomic dimensions, if heat dissipation does not stop improvement before that limit is reached.
There is absolutely no reason to expect the specific energy of batteries to increase in any regular fashion. Indeed, most improvement in battery performance has come about from changing chemistry rather than from ongoing improvement within a given chemistry, which has been incremental. I believe this is the “different time scales” referred to by the author. There has been improvement in battery specific energy from lead-acid through nickel-metal-hydride to the present leader, lithium-ion chemistry. Improvement, however, has been far from “doubling every 22 years.” The first sentence of the reference provided by the author, Zu and Li (2011), is as follows: “The average increase in the rate of the energy density of secondary batteries has been about 3% in the past 60 years.” This statement indicates that the specific energy of batteries has not quite doubled in the past 60 years.
Get our daily investorintel update
A recent EPRI report [Arshad Mansoor, EPRI, January 1, 2013] shows specific energy data from 1860 through the present, over five battery technologies. This report shows a doubling of specific energy over a 60-year period, roughly consistent with Zu and Li. This is far indeed from doubling many time periods of 22 years.
I leave it to the author to show data with specific energy doubling over several 22-year periods, which would indicate exponential growth (a “Moore’s Law” behavior). I maintain my view that there is no multi-period doubling of the specific energy of batteries, which would indicate exponential growth.
I suggest that particular care be used in citing Zu and Li. Figure 1 shows development of primary batteries (batteries which are not rechargeable), thus not relevant to transportation or many other applications. Figure 2 shows development of secondary (rechargeable) batteries, but includes data for Na/S and other batteries which are far from ready for use in the real world, as there are major issues to their implementation. Thus the most reliable quote from the paper is in the abstract: “The average increase in the rate of the energy density of secondary batteries has been about 3% in the past 60 years.”
Given the need to show at least 4 or 5 periods of improvement in specific energy density, and taking 60 years as an approximate doubling time (the real doubling time is greater than 60 years), something like 240-300 years of doubling every 60 years would be required to show exponential growth. This is longer than the time since the first useful battery, lead-acid, was first invented.
Thus there is no evidence at all that there is a “Moore’s Law” for batteries, as I stated in my 2013 paper, over any time period since practical batteries were first invented.
I’ll eat a bit of humble pie before moving to further thoughts on battery technology and transport, here goes…
I accept your comprehensive reply that there is no “Moore’s law” for batteries and I’ll retract my statement to the contrary unreservedly. I had also assumed the readership to be an informed audience that understood the empirical nature of “Moore’s law”. Your description of it as a self-fulfilling prophecy rather than a fundamental law of nature has admirable clarity that I should have employed.
You were quite right to pick up on the fact that simply doubling over a few time periods is not exponential growth.
Regarding the 22 year doubling period, I’ll happily share how that was derived from the Zu and Li (2011) paper. I do also take note of the care you advise in citing this paper. I’m very interested that you see the energy density increasing over significantly longer timescales (60 years rather than 22) and the EPRI report is certainly something I’ll pay attention to. Perhaps we can correspond outside of this column. My motivation here is continuous learning.
Enough humble pie.
This has made me think a bit more, and I’d welcome further views
We do agree that battery technology development is not progressing at an impressive pace in terms of the amount of energy a battery can store and release from a given volume or mass from the point of view of energy storage for transport.
If batteries are used for energy storage in transport they will have to be rather large and heavy devices and this is not going to change unless some unexpected breakthrough takes us all by surprise.
Given this limitation, it would seem that electric vehicles of the future have two possible lines of development:
- One could be to make batteries that can charge rapidly and frequently. This implies a well-distributed infrastructure of charging points built into the road network. My guess is that batteries with this capability need to have a very large surface area and a structural design that allows fast access to this surface to accept and release the charge.
- The other alternative for electric vehicles would be to store the energy in another form that has a high energy density and convert this to electricity on board the vehicle.
Many of the automobile manufacturers are already producing hybrid cars with gasoline or diesel engines that power on board electric generators. I am more aware of these types of vehicles on the road than electric vehicle charging points so perhaps this indicates the likely development path for electric vehicles of the future. I’d be interested in your thoughts too…
Adrian Nixon is a Senior Editor at InvestorIntel. He began his career as a scientist and is a Chartered Chemist and Member of the Royal ... <Read more about Adrian Nixon>