We don’t have time to respond on this website to every new
idea that comes along, but WindFuels is all about efficient energy
storage, so we’ll say a few things here about ultra-capacitors,
batteries, and the technology EEStor has been hyping.
The advances in porous-carbon-electrode electric-double-layer
ultra-capacitors over the past five years have been nothing short
of astounding, and there have been rumors circulating since at
least 2003 that another major breakthrough in ultra-capacitors
would soon be coming from some outsiders in Texas. Those rumors
swelled to a crescendo in August 2009, as some big investors jumped
into EEStor, even though no real data or sound theoretical explanation
is available.
Before looking a little more
closely at the EEStor story, it is useful to put their claim
into context. First of all, the EEStor
patents are not about ultra-capacitors (or super-capacitors),
as these words have been used over the past decade. Rather, the
EEStor
patents are about multi-layer ceramic (MLC) capacitors.
MLC capacitors, many using
various formulations of barium titanates, have been highly developed
over the past five decades by dozens
of large companies around the world. Hundreds of billions have
been sold. (For example, there are dozens in every cell phone,
and hundreds in every computer.) The chances of a breakthrough
that will allow a factor of 1000 increase in energy density and
a factor of 1000 reduction
in cost per energy stored, seem extremely remote.
The table below presents some useful reference data on energy
storage devices as of late 2009, listed in order from most to least
expensive in terms of $/kWhr.
The numbers in the last column have limited validity but are useful
for rough estimates of lifetime costs under some conditions. Capacitors
have historically been used for energy storage on a time scale
of milliseconds or less, so their lifetime in deep-discharge DC
applications is not well known.
The data shown in the
table for the lithium ion battery is based on oft-heard estimates
of the battery for
the GM-Volt. However,
some financial sales data from A123 suggest their
battery manufacturing cost has been $2000/kWhr. The data for
carbon-lead-acid batteries
is based on projections from Axion, a current leader in the field.
The data for the “conventional” lead-acid battery
is from EnerSys.
Only the last
three in the above table are competitive for most purposes,
though ultra-capacitors and flywheels compete in high-cycle
applications.
EEStor. So
let’s return now to the EEStor capacitor story.
The figure here illustrates a typical MLC capacitor – and
it is exactly what EEStor is claiming to have improved by four
or five orders of magnitude in both cost and energy density.
The common MLC capacitor of the past 4 decades. |
It really is trivial to improve on energy density
(at least in small capacitors) by a factor of 10 (perhaps even
a factor of 30) by simply not requiring the device to sustain
high voltages at temperatures above 70C or lifetime more than
a few hours. (MLCs are typically rated for use up to 125C and
lifetime greater than 20,000 hrs.) The challenge increases steadily
with size, both because breakdown is a somewhat random process,
even without defects, and because the probability of a defect
increases with size.
EEStor has done a brilliant sales job
by trying to focus the discussion on dielectric constant while
ignoring the subject
of E-fields
at the particle interfaces. They eliminate the negative voltage
dependence normally seen for the dielectric constant in titanates
by imbedding small poled particles in a soft polymeric matrix.
They are expecting to sustain mean E fields of 5E8 V/m for dielectric
thicknesses of 10 microns. Experience says they won’t come
within a factor of 10, and energy density is quadratic with E
field.
The composite dielectric proposed by EEStor (alumina-insulated
barium titanate particles in a polymeric matrix) will probably
permit somewhat higher energy density in MLCs below 70C compared
to conventional dielectric formulations. Some test results are
ambiguously reported in their second patent (7,466,536) that
suggest they may have achieved energy density two orders of magnitude
beyond that of conventional MLC capacitors for brief tests. However,
lifetime may be reduced by a factor of 1000, and cost of their
process seems likely to be much greater than for conventional
processes.
A few more technical details can help
to shed some light on their claims. A typical 1 µF, 100 V, MLC (barium titanate) capacitor
will have dielectric thickness of about 20 microns. If the operating
temperature is limited to 70C and ramp rates are limited to values
realistic for vehicle applications, it could have lifetime of
several hundred hours at 300 V. Extrapolating from data reported
by AVX, one would expect 16% of 1 µF, 100 V MLC capacitors with
25-micron dielectric thickness to fail in 1000 hrs at 200 V at
70C. (Because of energy losses during use, it will be hard to
keep the internal temperature lower than 70C on a hot day.)
EEStor reported tests at up to 5000 V (reportedly at 85C) with
dielectric thickness of about 10 microns. They also reported
mean breakdown E fields of 5.6E8 V/m. However, they apparently
used 1 micron thickness in their energy-density calculations.
Standard correlations typically give
an order of magnitude decrease in lifetime for a factor of
two increase in voltage at constant
temperature. It is not completely unreasonable to expect their
composite dielectric to withstand E fields as much as three times
those of sintered barium titanates below 70C for the same thickness
(as we argue later) with only a 20% loss in dielectric constant.
However, dielectric strength decreases almost as the square root
of thickness. (Yes, that’s true, even though most elementary
physics textbook authors don’t know that.)
The typical MLC achieves 0.007 Whr/kg for 250 V, 60-micron dielectric,
20,000 hr, 70C. If EEStor achieved 3500 V at 10 microns with
only a 20% loss in dielectric constant, their energy density
could have approached 8 Whr/kg. However, prior experience with
sintered barium titanate would put the limit at 130 V for 1000
hr lifetime at 70C for 10 microns. If they have achieved a factor
of three improvement in breakdown field (which seems plausible
for a complex composite), their energy density could be 0.1 Whr/kg
with a semi-acceptable lifetime.
There are sound reasons for doubting that more than a factor
of three increase in mean E field is possible over the current
state of the art. They note that the mean E fields they report
have been reported by others in very thin films of alumina. Indeed,
that is true. Breakdown fields in 20-micron films of good dielectrics
at 60 Hz are often 7E7 V/m, and in 0.2-micron films breakdown
may be 5E8 V/m. (In the 5-nm barrier in 2.3-V electric double
layer capacitors, breakdown may be 1E9 V/m.) The problem is that
during charge and discharge, the peak E-fields in the 10-nm alumina
coatings on the 1-micron titanate particles near the contact
points in their composite can exceed the mean static fields by
more than a factor of 50. If the coatings can tolerate peak E
fields of 2E9 V/m, the mean E field would be limited to 4E7 V/m,
or 400 V for a 10-micron dielectric, as assumed in the previous
estimate.
Some of the above estimates are based on 60 Hz data, but there
is not much difference in the DC data when it comes to voltage
breakdown.
EEStor plans to combine 30,000 1-mF capacitors in parallel to
make a 30-F, 3500-V capacitor. The largest MLC capacitors currently
available have energy storage about 0.002% that of the individual
capacitors they plan to produce.
So the bottom line is that they may be able to achieve 0.1 Whr/kg
(about 2% that of current ultra-capacitors) with 1000 hr lifetime
(also about 2% that of current capacitors), and they are likely
to be 1000 times more expensive than the competition.
References:
One of the few, sound discussions of how an ultracapacitor works:
http://www.mpoweruk.com/supercaps.htm
GreenTechMedia: Best source of recent technology progress:
http://www.greentechmedia.com/search/results/4aee4efb8179f0c0b41530cb954dd45d
A123 lithium-ion batteries:
http://www.greentechmedia.com/green-light/post/the-225m-ipo-roadshow-begins-a123-aone/
advanced lead-carbon batteries
http://www.greentechmedia.com/articles/read/axions-lead-carbon-batteries-sweet-spot-for-micro-hybrid-vehicles
US Patent 7466536, EEStor.
BS Rawal, NH Chan, “Conduction and Failure Mechanisms
in Barium Titanate Based Ceramics under Highly Accelerated Conditions”,
AVX Corporation, Myrtle Beach, SC
http://avx.com/docs/techinfo/barium.pdf
AVX: http://www.avx.com/docs/techinfo/c4interc.pdf
http://bariumtitanate.blogspot.com/2002/07/if-fascination-of-eestors.html
http://earth2tech.com/2009/09/04/how-risky-bets-like-startup-eestor-lure-political-backers
http://en.wikipedia.org/wiki/Grid_energy_storage
http://www.enersys.com/products.asp
http://www.modenergy.com/DS-RKU100-001G%2023in%20rackmount%20data%20sheet.pdf
http://www.geocities.com/CapeCanaveral/Lab/8679/battery.html
http://www.eere.energy.gov/de/cs_energy_storage.html
See http://dotyenergy.com/Markets/CAES.htm for analysis of compressed
air energy storage.
Energy Storage Association. Excellent overview of 2002 technology:
http://www.electricitystorage.org/site/technologies/
Field concentrations at dielectric discontinuities:
http://www.dotynmr.com/PDF/1981_JMR_Doty_DT_Solids_Probes.pdf
