breakthrough energy ventures, part 2
=companies =startups =energy =analysis
Let's take a quick look at some more companies in the portfolio of Breakthrough Energy Ventures. I'll continue going in alphabetical order.
Commonwealth Fusion Systems
A fusion startup, making a tokamak using different superconductors than
ITER to get higher field strength. I have a standard list of problems with
fusion power:
1) Anything
other than D-T fusion would be much more difficult. D-T fusion makes lots of
neutrons, which will damage the reactor and make everything around it
radioactive.
2) Even if you can get tokamaks to work, fusion is a
more expensive heat source than fission, and power plants using nuclear
fission are too expensive.
3) Plasma with fusion happening has
MHD instabilities. The sun has solar flares, and fusion reactors also
have "plasma excursions" that remain unsolved. This seems very difficult to
solve.
4) Not applicable
here, but various other reactor designs with non-Maxwellian plasma aren't
workable, per
this classic paper.
CubicPV
A startup making
perovskite solar cells (Hunt Perovskite Technologies) and a startup making
wafers directly from molten silicon (1366 Technologies) merged, and CubicPV
is now trying to do both. These are unrelated technologies but CubicPV wants
to also make
tandem
solar cells using both types stacked.
1366 Technologies got a
$150M loan guarantee in 2011 but decided not to use it. Their big idea was
using something like the float glass process for silicon.
Perovskite
photovoltaics are interesting, because they can be deposited from
solutions, but I don't see them being competitive
with monocrystalline Si, mainly due to the stability issues of organic
components in the ones with good performance. Also, the current best ones contain a lot of lead,
which I consider unacceptable.
A thin layer of water on a windshield
beads up into droplets: there's a minimum thickness to get a smooth sheet of
water, which depends on the surface tension of water and the strength of its
interaction with the windshield. Float glass also has a minimum thickness.
Molten silicon has high surface tension, and the stuff it sticks to well
also tends to dissolve in it, and dissolved impurities in silicon wafers are
really bad. People figured out how to cut thin wafers from ingots by using
small wires with embedded diamonds, so the direct wafer approach doesn't seem
competitive with using monocrystalline Si ingots.
As for tandem solar panels, it would
make more sense to use CdTe + silicon, but that's been too expensive so far.
However, the non-viability of these approaches vs monocrystalline Si is
a fairly difficult tech evaluation, and it was correct for institutions to
invest some amount in their development, since it's not plausible for them
to make that call accurately. CubicPV certainly isn't something obviously
dumb like Solyndra or Theranos, and while I think the company is doomed,
they probably have smarter people than the companies doing obviously-flawed
things, so people should hire them.
Silicon is cheap; maybe people
should just be making more silicon in America instead of relying on
China...? Currently, solar panel supports are more expensive than the actual
panels, and the cost of inverters is comparable to the cost of solar panels,
so I'd actually like to see more work on those areas rather than the
photovoltaics right now. Alternatives to indium tin oxide (the transparent
but conductive coating) also seem like something worth working on.
C-Zero
This company is
converting methane to hydrogen and solid carbon. That's fairly easy: usually
process designers are trying to avoid that happening during methane
reforming by adding enough steam, mostly because the solid carbon builds up
on stuff (coking) and needs to be burned off. It's also less valuable if
collected: usually people (mostly people in China) are trying to convert
coal to syngas, not intentionally produce less syngas to get something like
coal.
I'd expect this to be
significantly more expensive than normal methane reforming, because of the
solid carbon handling issues. The resulting carbon would just be buried,
which makes less sense than not digging up coal, which is still a
thing that's done a lot.
I'm sure it's
possible to do this, but it seems significantly more expensive per ton
CO2 emitted than using biomass for stuff.
Dandelion Energy
This is a
ground
source heat pump startup. That's not a new technology, and it works well
but has generally been more expensive overall than air source heat pump
systems.
I don't see anything technologically new about this company;
they seem more focused on the finance side.
DMC Biotechnologies
This
company is doing metabolic engineering for production of amino acids. Amino
acids from fermentation are used in animal feed and sometimes for other
stuff, so sure, that makes sense. Can they do that slightly better than the existing
companies in that area? Well, maybe. They seem competent enough with the
biology part, at least.
Ecocem
This company is
blending blast furnace slag with Portland cement. This is a very normal
thing to do with blast furnace slag, and there are a bunch of companies that
do that. Nothing new here.
Electric Hydrogen
Making
hydrogen by electrolysis is >2x too expensive and this company isn't going
to change that. This is particularly absurd as a plan for utilizing
unreliable solar power, because you don't want to run electrolyzers or stuff
using hydrogen for 6 hours a day. Anything but basic research on this topic
is premature at this point.
At least that's better than trying to
directly convert CO2 to chemicals, I guess.
enVerid
This company sells
HEPA air filters with UVC lamps. That makes sense and lots of people do that
now.
They also sell sorbents for air treatment, basically activated
carbon plus something that absorbs CO2. Activated carbon is more expensive
and less important than HEPA filters, but it's used sometimes. Absorbing CO2
indoors is their special new thing, but that just seems silly to me: the
amount to be absorbed is much higher. I can't see that being viable vs more
air flow from outside with heat recovery ventilators and HEPA filters. Maybe
it makes sense as a luxury thing for really rich companies and
individuals...?
ESS, Inc
ESS is making
iron flow batteries, iron metal <-> FeCl2 <-> FeCl3 in water. They need anion exchange
membranes, which are expensive.
This approach was first
published in 1981. It has issues with hydrogen generation and ion balancing,
and ESS is presumably adding a separate rebalancing system.
Flow
batteries are one of the more plausible types of grid energy storage, and
are sometimes used for power backup systems. This seems better than Form
Energy's approach. However, the membranes are too expensive, and ESS has no
real path towards reducing their cost sufficiently. So, this isn't viable
for grid energy storage, and it's probably more expensive than LiFePO4
batteries for backup power...but sure, this is a legitimate energy-related
startup investment, I guess. Good job, Breakthrough Energy Ventures.
Low voltage is a big problem for flow batteries, because at a given power
level, losses are proportional to resistance / voltage^2.
There have been a number of failed flow battery startups; I think
the only commercially available systems today are vanadium and zinc-bromine.
EnerVault tried to commercialize iron-chromium,
and it went bankrupt in 2015.
There are some better flow batteries
designs, and with those, solar +
wind + flow batteries with a national HVDC grid would still be more expensive than
natural gas + solar + wind in the USA, but the premium should be <$0.10/kWh,
less than the premium California residents are paying today for PG&E
corruption, but probably more per ton CO2 mitigated than doing certain stuff
with biomass.
Fervo Energy
This is a
geothermal power company. Their supposed advantages are:
- horizontal
drilling
- fiber optic sensors
- cloud networking
- machine
learning
The companies that have been drilling for oil for 50 years are actually pretty competent, and geothermal drilling seems to make more sense as a side thing for such companies, unless you're trying to do something new like drilling with microwaves or using diamond bits in really hot rock - neither of which would work out for you very well.
Form Energy
They're making
iron-air batteries, which is the same as zinc-air but with zinc replaced by
iron, so they're using positive electrodes from a company that's worked on
zinc-air batteries. Those are expensive. I believe they're currently more
expensive than lithium-ion per watt, probably more than $300/kW.
Iron-air and zinc-air batteries also use ion (probably anion) exchange
membranes, which Li-ion doesn't require. That's the other part that's
expensive per watt, but that's cheaper than the positive electrodes.
Form Energy's "solution" is to charge and discharge really slowly to reduce
those per-watt costs relative to capacity. Iron is used instead of zinc to
avoid dendritic deposition problems with zinc, but there are significant
disadvantages: iron-air gives even lower efficiency, about 50%, and it's
even heavier than zinc-air.
Bill Gates has invested in a number of
grid storage startups, including LightSail and Malta. His goal is
preparation for a future using mainly solar power with energy storage. He
set a threshold of 60% plausible storage efficiency, but that's not always
reached - for example, Malta says they'll reach that but I don't think they
will.
Such a future won't come in time for these startups to survive,
so I believe the plan is accumulation of intellectual property from them. In
the medium term, low efficiency and slow discharge isn't economically
viable.
The cycle life for iron-air is generally at best 2000 cycles.
With wholesale electricity prices around 3-4 cents/kWh, low efficiency, and
slow discharge rates, it's hard to argue for a profit over 4 cents/kWh.
Considering the cost of power electronics, and interest on borrowing, they'd
obviously need to reach $70/kWh to even plausibly be useful. That's not
impossible if discharge takes several hours, but it implies thick iron
electrodes, and I'm concerned about that reducing cycle life due to
differences in iron deposition rate - but I'd have to see their test data
regarding that. Form Energy is claiming they'll reach $20/kWh but they
won't.
For comparison, lithium-ion lasts around 1000 cycles when
derated to 80%, and Tesla Megapacks are ~$350/kWh, which is obviously at
least $0.35/kWh of grid storage. Due to a mandate for battery storage from
the California gov, it looks like PG&E will be spending ~$400 billion on
Megapack-based battery storage at the Moss Landing site - enough to power
California for 2.5 minutes. So, government mandates can always change
things.
H2Pro
This is a company
trying to do an
unusual type of water electrolysis. The oxygen goes into a nickel oxide
electrode, which is heated afterwards to release oxygen. This means you need
a lot of electrode surface area, and a lot of heat transfer. Their process
uses a slightly lower voltage than typical water electrolysis, but I can't
see it being cheaper than that due to the costs of the electrodes and heat
management.