March 10, 2012

Lunar Silicon vs Helium 3

What if a lunar base could produce and deliver waste silicon cheaper than coal on Earth? The economics of massive lunar exports

A guest article by Joseph Friedlander
There was a lot of talk stirred up lately by Newt Gingrich's lunar base speech, which Brian referred to in a post here.

Many people have been dragging out the corpse of Helium 3, whch was really popular around 1990 or so, to be propped up for some one-sided parlor conversation about an economy on Earth powered by Helium 3.

Yes, the Moon's powdery regolith, over much time, has adsorbed (not absorbed, adhered to the surface of the fine grains) hydrogen, nitrogen and helium from the solar wind, among them Helium 3 atoms.

Yes, this could (by moving vast amounts of regolith and heating it) be mostly captured. We are talking a couple million tons of regolith (it does not go deep, only the top meter or so) per square kilometer, and gardening millions of square kilometers over time (trillions of tons of regolith heated to extract the Helium-3)

Yes, theoretically fusion reactors could be developed (D-T is easier) that MIGHT someday use Helium 3.

Yes, the lunar resource would possibly last up to a single century (The outer planets would presumably be mineable after that time)

But I find this talk unconvincing-- it's what I call "shoehorning a dream". You want to fit this dream into this not really compatible bunch of facts, and by golly you have got a shoehorn you are determined to use-- so in it goes, whether it wants to fit or not.

What Helium 3 advocates want really is a conveniently labelled bulletproof Lunar export that is vital to the economy yet light enough to fit plausible traditional traffic models (using rocket transport) yet expensive enough to add up to huge bucks and a huge slice of the economy. There is huge energy demand so you could mine and market a lot of Helium 3.

IF..IF..IF the reactors existed, which they do not. But lunar diamonds, for example, would crash in price if you tried to sell into Earth's tiny (and frankly cartelled to death) diamond market.

Now this is frankly a totally speculative post, but let us try the opposite of that, a huge lunar export which this article will admit frankly depends on a massive lunar industrial infrastructure that does not exist. But hey, the Helium 3 reactors don't exist either, so fair's fair.

Let us suppose we had:

1.Gigantic moonbase industrial complex
2.Huge (thousands of square kilometers, maybe tens of thousands growing to millions) of mylar or foil reflectors (12 tons per square kilometer) to heat regolith
3.The ability to vaporize regolith and rocks to plasma
4.The ability to magnetically deflect the elements so oxygen is diverted from the metals and metalloids (ideally to capture it-- see my post on the Friedlander Cold Crown for capturing lunar oxygen on NBF.)

5.The ability to condense each stream and pile up 'fallout' of each relatively pure element
6.The ability to scoop up this element and redistill it for great purity in vacuum
7.The ability to form it into any configuration we need (for example, casting entry body, making fibers for cables or tethers
8.The engineering ability to build power stations, big electric motors and launchers so cheap to use that power is the dominant cost, not amortization charges.
9.We assume for this article penny a kilowatt electric power on the moon, as we have covered before concerning the Criswell plan for lunar surface power. (Google criswell+lunar+solar+power) And far cheaper thermal solar power.
10.And all this can be done so cheaply that the speculation in this article is possible.

Well, that probably encapsulates a few space industrial revolutions in a short list, so lets go on.

The key thing that is easy to miss in that list is plasma reduction of the lunar regolith.

Around 1969-1971 there was an idea called "The Fusion Torch" by Eastlund and Gough that involved fusion vaporizing rocks or garbage and pretty much doing what I listed above--separating compounds, dissociating and separating them, and reclaiming the elements.
Google "energy waste and the fusion torch"

a sample picture is here

This is the report about the fusion torch-- warning, LaRouche political site, no endorsement intended, but they are hosting data on the fusion torch. papers

The key takeaway from this is that temperatures above 4000 K can basically vaporize all rocks, 6000K (Solar surface temperature) breaks up nearly all compounds except things like silicon fluoride and cyanogen, 8000K breaks up basically all compounds. Basically 16000K will ionize everything but extensive ionization occurs in some species between 4 and 6 thousand K.

The point of ionization of course is that you can deflect different species selectively like a mass spectrometer. It gives huge control.

Note that this is a matter of a bell curve-- above 4000K some of almost every compound will be dissociating, the question here is where on the bell curve you draw the line after which NO compounds remain and all are elements dancing in a plasma.

I believe based upon study of the problem that around 4000K enough will be dissociated that repeated passes will yield commercial amounts of the elements in question. With solar power (at a guess, around an intensity C of 100,000 times normal sunlight) that temperature would be achievable for a small fraction of the heated material. Some fraction might be ionizable or deflectable in other ways. Many industrial processes are like that, going for an achievable 1% of a flowstream reacting and separating that rather than an unachievable one step 100% ideal.

The idea is -- vaporize--dissociate--deflect-- separate--repeat.

It could also be that merely chemical effects are enough, that more volatile species of oxide will evaporate, lesser will remain and we separate that way.
If we can do it thermally rather than through electrical equipment we save hugely on infrastructure. But if we can't avoid massive electricity use in the fusion torch like process we want to develop solar lunar power for export anyway, so...let's go.

Notice that we are not embracing the usual vision of a dozen white modules on grey powder visited once or twice a year by ultra expensive lunar modules with a few tons of export capacity, but huge refineries with something with multi thousand ton electro catapult cargo launches to Earth every hour even early on. I may write an article on the logistics of this later on but this is not it.

Current experience is that say $2 billion will let you send one ton one time from the Lunar surface to Earth, in this article we assume that will fall by many orders of magnitude (Moore's Law for Lunar surface launch costs) so a mass driver like device of a nature I do not have to specify here allows something to be shot to earth for a penny or two a kilo. (At a penny a kilowatt-hour the power cost would be about 2/3 of 1 cent per kilo). So a factor of a hundred billion times cheaper. (200 billion pennies is $2 billion dollars. Count 'em.)
This only would work with massive throughput to amortize costs
...fortunately Lunar launch direct to Earth is often --because the Moon is tidally locked--easier than many people assume, and far easier than to orbital destinations. We assume the scale and logistics of this article as a given just to get to the main discussion: What is there on the Moon that can plausibly be exported?

1.Microwaved electricity (once generated) via the Criswell Lunar surface power plan-possibly involving sprayformed solar cells directly on lunar surface in vacuum and sprayformed microwave rectennae.

2.Moon dirt --regolith- and the most common elements therein once separated
3.Lunar Basalt, quartz and silicon fiber if you can make it
4.Lunar metals (most common and marketable: Iron, Aluminum, Magnesium, Titanium. Scarcer but plausible-- phosphorous and chromium and manganese)
5.Later on formed metal products (this involves much more industry than at the beginning)

But when you make such a list you start doing a resource assessment --

Here are typical contents of the lunar regolith by main regions:

Chemical composition of the lunar surface regolith (derived from crustal rocks)

Compound        Formula   Composition (wt %)
                Maria Highlands
silica          SiO2       45.4%   45.5%
alumina         Al2O3      14.9%   24.0%
lime            CaO        11.8%   15.9%
iron(II) oxide  FeO    14.1%    5.9%
magnesia        MgO         9.2%    7.5%
titanium dioxide TiO2       3.9%    0.6%
sodium oxide    Na2O        0.6%    0.6%
Total                      99.9%  100.0%

Wikipedia on the moon

Note that this is one set of tables on a global average, specific sites could for example double the titanium (and note also these are expressed as oxides, so for example that's not 4% titanium but more like the table above.--But in the Sea of Tranquility, for example, you could probably get 4% in random regolith. If you can export from multiple sites, you can often find 10% iron, in a separate place 4% titanium, perhaps 12%+ aluminum, and at another site as much magnesium--using totally different regolith types.)

Keep in mind that by weight oxygen is about 40 % of the lunar regolith and silicon and calcium together around 30 % and you will see that potential export metals are only about a quarter of regolith weight.

So I started thinking-- could silicon be a plausible Lunar export?
(People ask about exporting lunar oxygen to future space industries, for massive space colonies, etc. I say build them first and we will talk about it. - You are assuming vast space infrastructure that exports to each other. I am saying now, how are you going to finance that--where are your customers? They don't exist yet. How are you going to bootstrap that?!)

This article is assuming a lunar industry can be built if able cheaply to meet demands on Earth which already exist--or can be plausibly booted up given real economics.
Yes you through away your gravitational advantage of already being in space--but until paying customers are in space you have no use for that gravitational advantage. (And there won't be many space customers until a big enough solar economy gets going. Most paying customers are and for the near future will be on Earth. Let's try to keep to one fantasy at a time.)

So the answer becomes what will Earth pay for and at what plausible price?
Notice what we are doing--not looking for the scarcest materials on the Moon but the commonest and using the natural advantages of sunlight, metal bearing rock in contact with vacuum, and low gravity to ease space launch (and the tidal locking of the Moon to greatly increase launch windows.) Not to mention that the Moon and the Earth constitute a united system whose members have short travel times between them with no long synodic waits.

So let's focus now in the rest of this article on what could be done with lunar silicon, which is 20% or more of most lunar regoliths--
1.It can be used on the lunar surface in spraybuilt solar cells for the Criswell lunar solar power plan.
2.It can be exported (once we have massive power generation and some sort of mass driver that does not depend on rockets.

Concentrating on silicon exports to Earth (Why? Because it's a symbol of the Lunar byproduct resources thought worthless--as in we have plenty of silicon on Earth and you are going to import it? The answer is, yes-- because the silicon on Earth is oxidized and the Lunar silicon can be freed with the supercheap solar energy possible only in space. It is at least possible --and this article assumes as part of its premise-- that this is at so cheap a price that literally it would pay to import freed elemental silicon from the Moon than to refine it from the sand near your house.

For this to work depends on transport cost-- we assume supercheap export and shipbuilding on the Moon-- (The shipbuilding will be covered in another article)
We do mention here that the metal of the cargo ships shot toward Earth could be used as seasteads or as scrap for breaking up; a huge industry in India breaks up ships today, but earth ships are very complicated, and these barges would be designed for easy disassembly. So let's say 10-20% the exported silicon weight is in scrap iron for the cargo barges.

Note that if we could capture the kinetic energy of an incoming cargo barge at 11 kilometers a second that would be a huge energy payoff: To export a kilogram from the Moon is about 60% of its' weight in TNT Something falling over the gravity hump toward the Earth though will generate 24 times this amount of energy (the Moon is at high potential relative to Earth)

The gravitational potential at the Earth's surface is –62.5 MJ kg–1 and at the orbit of the moon it is about -0.5 MJ/kg

In fact, as Paul Birch pointed out long ago, importing all the metals Earth uses from the Moon would also import all the energy we use in that case. 1% the mass of the Moon imported could power the Earth for hundreds of millions of years, simply by tapping the potential energy. (The orbit of the Moon would be visibly affected after say a hundred million years, but the gravity of the Earth would increase by way less than a thousandth)

But we assume no kinetic capture energy available here, just selling the bulk materials on the ground. Let's explore the economics of lunar silicon exports.

The rule of thumb is, the cost of coal is a good model for the selling price of lunar silicon. (Because it is absolutely proven that there is a market for a fuel and deoxidizing agent of that kind at that price) The question is, can we produce and export it for that price and a little cheaper.

What would such a lunar silicon export market be like?
1.A small market at first for ferrosilicon (mixture of iron and silicon) and elemental silicon at high prices.

Often around $2 a kilo of silicon.$2000 per ton. China, the leading supplier of elemental silicon, providing 4.6 million tones

2.A comparable market for specialized things like silicon fiber and pure quartz fiber
3.A much larger market for elemental silicon to be burned as solid fuel like coal. This can be for fuel or for hydrogen generation via the reaction. This could be billions of tons but at a mere $80 to $100 a ton.
(probably in a fluidized bed reactor) Si + H2O = SiO2 + 2H2
So eight tons of silicon would free a ton of hydrogen

Here the reaction is similar to the water-gas shift reaction

CO + H2O= CO2 + H2

Hydrogen production by shift reaction: I have seen a quote of $1,900 per metric ton; I have seen quotes for about a third that, years ago. Often natural gas is stripped of hydrogen, a more expensive feedstock. Whereas straight coal burning might be worth $80 a ton to the silicon supplier hydrogen production might be worth perhaps twice that (or not) if the reactivity is more favorable. This is also speculative.

Present U.S. use of hydrogen for hydrocracking is roughly 4 million metric tons per year (4 MMT/yr). It is estimated that 37.7 MMT/yr of hydrogen would be sufficient to convert enough domestic coal to liquid fuels to end U.S. dependence on foreign oil importation and less than half this figure to end dependence on Middle East oil.

A sample interesting device to use silicon to make hydrogen is here--it looks like a small scale gimmick, but it illustrates the possibility of a silicon fuel cell 76 kilograms of elemental silicon +100 liters water yields 11 kilograms of hydrogen and 163 kilograms of sand and 7.5 kw fuel cell output at 50pct efficiency-- 180 kilowatt hours of electricity is the claim

In real life I am far more confident that hydrogen liberation and hydrogenation of coal and tars would create usable liquid fuels to fuel the vast infrastructure using hydrocarbons, which we already have.

What about EROEI, you ask?

Let us suppose that to heat 1 kg to 4000K takes around 1.1 kilowatt hours. Let us suppose we cycle this 20 times in the various refinings between raw rock and distilled species of metals. (Launch to Earth is negligible 1 kilowatt hour and less) Everything total would be about 30 kilowatt hours per kilogram-- yet silicon burned yields only around 8 kilowatt hours per kilogram. So EROEI of about .25 (one quarter of unity) or less, perhaps much less. It could easily end up being 100 kilowatt hours in and 8 out. However this basically is bottling sunshine on the Moon and exporting it to earth along with the silicon.
This is outside the Earth system, so it does not get charged to Earth resources. That is one of the advantage of a space colonial economy-- it brings new resources to a table and eases the pressure on the mother planet. Yes, the sun is losing an enormous amount of energy a day (4 million tons of energy a second, an impressive asteroid's worth a day--and remember pure energy is an atomic bomb's yield per paper clip weight of energy) but if the Earth vanished tomorrow it would keep on doing it. It's not our dime, friend.

This is a perfect example of engineering (energy flow) EROEI vs. financial EROEI-- the sun shines anyway and it ultimately all goes to waste. If with a foil mirror in vacuum (100 micron aluminum) weighing 270 tons per square kilometer costs $1 million you can capture 4000 gigawatt hours (4 billion kilowatt hours) per year with a cost per kilowatt hour (thermally) of a fortieth of a cent. So say we use 30 thermal kilowatt hours, that is .8 cents a kilogram or 80 a ton just for power costs. (If we assume 4 year payback then $20 a ton just for power costs. Assuming a multiplier of 4 beyond power costs--$80 a ton. Just about comparable with coal (given the greater fuel value).

Let us imagine the landing barges, a few thousand tons and tens of meters across, made of metals to market, aluminum, titanium, iron, magnesium. 80-90% of the cargo would be silicon. Remember--you cant dissociate regolith without producing as much silicon as the good metals. Its a waste product which if sold at marginal prices can defray mining and exporting costs for your more profitable metals.

World market
Steel production per year- 1500 megatons $.43 scrap value typical
aluminum production per year- 40 megatons $1 kilo typical
45 million tonnes, over 90% of which is used in the manufacture of aluminium metal. say 50c a kilo typical
(obviously if you displace aluminum metal on the market those companies do not need to buy the raw materials for aluminum).
magnesium production per year- 600 kilotons $1 kilo typical
titanium production per year- 100 kilotons $10 kilo typical (20 times this in oxides)

Coal, 7 billion tons a year, say 8 cents a kilo typical ($80 ton)
has many important uses worldwide. The most significant uses are in electricity generation, steel production, cement manufacturing and as a liquid fuel. Around 6.1 billion tonnes of hard coal were used worldwide last year and 1 billion tonnes of brown coal

Coal prices (for comparison) lately typically $70--80 to $130-140

What would the natural markets of elemental silicon fuel be?

Power stations or chemical plants near the sea (we assume water landing but inland landing at any inhabited height is possible) near markets for large amounts of construction materials, fill, and other sand markets.
Note that when you burn silicon you don't add CO2 to the atmosphere, you take oxygen out and get twice the weight in pure quartz sand, which can be used, for construction, glass making, and seaside landfill to make new real estate that you can sell.

It also by definition is clean fill; never having had the organic or industrial waste poisons which many Earth sands have absorbed near major cities.

Cement makers could burn silicon to make cement and then sell the sand and cement for use in concrete. For example, an economy the size of New York City or Israel (in a developed country) can easily use 15 million tons of coal a year just for heat and power. This amount of silicon, burned into sand, would yield a square kilometer of new landfill, ~12-14 meters deep. So over decades large archipelagoes of new land along the coast might be made.

In the USA alone people pay for around 800 million tons of sand and gravel yearly; that's about the equivalent of burning 400 million tons of lunar silicon.

Oil refineries and other chemical plants that want gaseous hydrogen to upgrade products. Metallic calcium from the moon if cheap enough would also have supplemental chemical uses in the metals refining industries (nitrogen removal) and even ammonia manufacture.

We can imagine power stations or chemical plants near New Orleans burning lunar silicon for local industries just to build up breakwaters against another Katrina. (Obviously cheap lunar-beamed power would compete with this, but we are computing material exports in this article as a thought experiment.)

A similar fill market in the Netherlands is obvious.
(And Tokyo Bay, and many coastal Chinese cities---)

What is the thermal value of silicon?
More than either bituminous coal and comparable to the 1/3 more energy intense (and rarer) anthracite coal and aluminum metal. (If you burn them)

What is the thermal value of bituminous coal?

The thermal energy of bituminous coal is approximately 6150 kilowatt-hour (kWh) per ton or around 24 megajoules kilogram enough to melt say 15.375 tons aluminum per ton of coal (400 kwh ton to melt)

(Incidentally I considered briefly importing molten lunar aluminum to sell both the metal and the heat, but coal is quite energy dense compared to molten metal 960 °C (1760 °F) is 0.39 kWh/kg of aluminum--as you can see a tiny fraction of coal's 6.1 kWh/kg energy intensity.

From Wikipedia's energy density article

Energy densities ignoring external components
This table lists energy densities of systems that require external components, such as oxidisers or a heat sink or source. These figures do not take into account the mass and volume of the required components as they are assumed to be freely available and present in the atmosphere. Such systems cannot be compared with self-contained systems.
Energy Densities Table - Energy Media Only
Storage type Specific energy (MJ/kg)

Coal, anthracite
Coal, bituminous
Coal, lignite[citation needed]

32 megajoules = 8.88888889 kilowatt hours per kilogram

What would be the upper limit of demand for lunar silicon fuel on Earth (present demand) if lunar beamed power was impossible but thrown and landed lunar fuel cargoes were easy?

Say about 25 billion tons lunar silicon year (less than 30 gigatons coal equivalent, ) over $2 trillion in fuel per year

A large part of this could be used to make hydrogen, in order to gasify and liquify coal into cleaner and usable fuels such as methane and synthetic diesel.
Fuel type Average power in TW[22]
1980 2004 2006
4.38 5.58 5.74
1.80 3.45 3.61
2.34 3.87 4.27
0.60 0.93 1.00
Nuclear power
0.25 0.91 0.93
Geothermal, wind,
solar energy, wood
0.02 0.13 0.16
Total 9.48 15.0 15.8
Source: The USA Energy Information Administration

There is one little catch to all this-- although the kinetic energy of lunar rock could last for hundreds of millions of years, the ability to burn silicon as coal would be far shorter lived till we had to give it up.

The Earth's atmosphere is 5 million gigatons

so 1 million gigatons of oxygen
so 10000 gigatons of lunar silicon burned up uses up as much oxygen, an arbitrary limit to this would be say the equivalent of 1% of the atmospheric oxygen (.21% of the whole atmosphere, 1/500 of the atmospheric pressure.

This would take 333 years at 30 gigatons a year.

The current realistic limit--if this happened today--would be 10 gigatons a year because you couldn't sell much more scrap iron than that (assuming 10-20% weight of payload as barge mass. That would be good for 1000 years.

And the likely limit is far less (it will take specialized or at least adapted burners and reactors to use the silicon, not everyone will switch). My bet, as noted above, is coastal locations that can use good fill—an idea originally brought forth by Paul Birch. But replacing say 25% of coal would still cut carbon emissions and help fund lunar industrial buildup. (Since the scrap iron sent down does not get refined using coke (roasted coal) that too cuts down carbon emissions).

Realistically, I can see this as an interim bootup industry for a few decades, but it will be as obsolete as whale oil once Lunar industry develops better (more finished) products. (And also as Lunar industry masters solving tapping the kinetic energy of lunar rock, beaming power, etc.
In any case, this is only an article outlining what COULD be not what WILL be.

We could make it work for far longer by importing liquid oxygen for the kinetic energy and then burning also imported silicon but long term I am sure this would be a transient business.

Boot up teardown schedule

Needless to say the first export products of lunar industry would be relatively light and valuable. But in refining large amounts of regolith to make aluminum metal (say) magnesium and titanium, iron will be produced for barge hulls, ceramics for heat shields, and scrap silicon will be produced. Eventually someone will get the bright idea of using it to fill out an underweight load and selling it for what it would be worth--at first as valuable metallurgical silicon at $2 a kilogram, then for what it will fetch.

Eventually the history of the commerce will take a bell curve like aspect on a chart, at first explosive growth because of great cheapness, then saturation, then decline (perhaps as yet cheaper beamed Lunar power and Earth based fusion torches replace the once-cheaper Lunar physical imports).

Economic history moves in cycles.

So what have we learned from this article?

1. With enough assumptions you can make anything fly (literally)
2. Industries have byproducts and will try to sell them
3. You sell the most valuable stuff first then the less valuable.
4. The best exports are what you have lots of easily accessible and a comparative advantage at.
5. A real export is to a real market, not a hoped for or imaginary market.
6. .Solar thermal power should be really really cheap on the Moon—half the time (daylight).
7. You can waste huge amounts of energy that cheap and still make money if you are exporting to a market where energy is dear such as Earth.
8. Exported lunar combustibles (refined metals and metalloids) though a huge waste of energy from the view of producing them might well be economical fuels.
9. They could certainly give a break to carbon emissions for an interval till we develop lunar broadcast power and other clean sources.
10. There is a limit to how long we can use them.
11. Tapping lunar solar power and beaming it to Earth and lunar kinetic power (the energy of incoming loads) makes a lot of sense over the long term.

The illustrative point of all this is that rather than taxing ourselves $100 trillion to build up an lunar industrial base with which to found a solar empire, we could buy the power we are going to use anyway from the Moon and let that pay for it.

If you liked this article, please give it a quick review on ycombinator or StumbleUpon. Thanks

Форма для связи


Email *

Message *