Wednesday, October 30, 2013

Krampus Wish List - Silicon for PV and Electronics


This post is in response to the archdruid's Krampus wish list post:
thearchdruidreport.blogspot.com/2013/01/a-wish-list-for-krampus.html
He wishes for various things to contribute to a sustainable way of life in a post-industrial world, and solicits concrete examples of workable solutions.

I've divided my contribution into three parts, the preparation of silicon wafers, photovoltaics (PV) from them with a discussion of other technologies, and then a bit about simple electronics.

Part 1 - Silicon for semiconductors (PV and transistors)

Silicon was first isolated as a pure element in 1823, though not until 1854 was it reduced in crystalline form (from electrolysis of molten salts containing silicon).  It was only in the 1940s that interest in it for its semiconductive properties began in earnest.

Worldwide, about 2 million tons of "metallurgical grade" silicon (mgSi) are produced these days, with another 6 million tons of silicon in the form of ferrosilicon for the steel industry. [1]
About 3/4 of the mgSi is used in the chemical industry to make silicones, most of the rest is an alloying element in aluminum. About 250,000 tons per years is purified into "electronic grade" silicon (egSi), and most of that goes to photovoltaics (PV). [4]
Only about 25,000 tons/year is used in the entire worldwide (non-PV) semiconductor industry. [5]

EgSi forms the basis of the world-wide electronics industry, with more than a trillion dollars in sales [2], as the semiconductor of choice for most logic, memory, analog and power devices, ceding only a few esoteric markets in radio frequency, high temperature encironments and optoelectronic devies such as LEDs to more esoteric semicondutors.

Why bother with silicon in a de-industrializing world?

First, PV is one of the few renewable sources of electricity that is scalable from a few watts which can charge a flashlight or run a simple radio, up to Gigawatts to run large industrial process, widely available (unlike hydropower or wind), and sustainable until the Sun ends life on Earth. [3]
Secondly, even fairly "low tech" semiconductors like power transistors, operational amplifiers, signal transistors, diodes and the like can add a great deal of ease and efficiency to life and economy, without being that expensive/complex to make.

Now obviously, if civilization devolves back to "cave man" status, we're not going to be sitting around using PV.  But if the level of sophistication is maintained that can make moderately complex steam engines (e.g. similar machines - compressors and the like),
hydroelectric plants and/or medium wind turbines (e.g. Megawatt power sources), then I think we can have PV and some simple electronics.

Silicon is not rare by any means, it is the 2nd most common element in the crust of the earth, the first being oxygen to which it is commonly combined.  We've pretty much all seen common sand, quartz crystals and the like - that's silicon dioxide, the raw material of silicon.

The processes for reducing and purification of silicon are not very complex.  There are two in use today, electric arc furnace and hydrochlorination, and the use of silicon tetrafluoride, a byproduct of the phosphate fertilizer industry.  The end result of both is a purified volitile silicon compound that is then reduced to pure polysilicon with heat.

_Electric_Arc_Furnace_and_Hydrochlorination_

The electric arc furnace (EAF) was demonstrated 1810 [6], the first commercial use was for calcium carbide 1888 [7]. Note that calcium carbide and water produce acetylene gas for carbide lamps [8] or welding/cutting torches [9], so there might be a complementary demand for EAFs for that in the future. An EAF is fairly simple, graphite electrode(s) create a big arc, either between themselves or to the furnace lining. The electrodes are currently made from petroleum coke or anthracite coal stuck together with coal tar pitch, but can be made from wood charcoal and tree pitch. For reasonable scale one needs about 1 MW of electricity, such as one would get from a medium hydroelectric plant or big wind turbine, or a middling big PV array.  (Niagra falls and similar places were the first sites of EAFs). But they can be made quite small if desired.  Current large ones are in the 100 MW range, producing many 10's of thousands of tonnes of mgSi per year.  A small 150 kW furnace at NETL makes roughly 5 kg of mgSI/hour. [30]

The basic recipe for silicon uses silica in the form or pure sand and/or crushed quarzite, added to an EAF along with chunks of carbon and wood chips.  Commonly today the carbon is from coal or coke, but wood charcoal works just as well.  The wood chips are a vital part, as the gaps between them provide a means for the carbon monoxide to rapidly escape the reaction, which helps keep the production of silicon high and lowers the production of silicon carbide. [10]  The desired reaction is:

        SiO2 + 2C = Si(l) + 2 CO(g)

The EAF is tapped occassionally to pour off the mgSi and any silicon carbide.  Note that silicon carbide will likely be needed as an abrasive in the cutting of silicon wafers.  Silicon carbide may also be synthesized in an EAF or other electric furnace, with no wood chips, just silica and carbon packed in.

The produced silicon is around 99% pure, but needs to be many more times pure for PV (say 99.9999% pure) and even more pure for electronic (99.9999999% - "nine nines").  This is done by reacting the mgSi with hydrochloric acid (HCl) vapor.

So, where to get the HCl?  The oldest method, known to the European alchemists, is to use sulfuric acid on common salt (sodium chloride). [11]  Sulfuric acid is easy, one burns sulfur or roasts sulphide ores, ideally then further oxydizes the SO2, and then washes the gas in water.  This is old stuff, and the industrial "lead chamber process" in 1746 essentially ignited the modern age of chemistry. [12].  Sulfur is fairly abundant, there are great stacks from desulphurization plants in various parts of the US and Canada, resulting from desulphurization processes, and the lack of economical use/transport. [13]  Sea salt is abundant in the sea, and there are many rock salt deposits left.  We can recycle the HCl, so long term needs are minimal.

A newer method is electrolysis of salt water.  Since one likely needs some sodium hydroxide ("lye") for other purposes, the chloralkali process would come in handy. [17]  This process began industrial scale in 1892.  All you need is pure NaCl brine, water and electricity to product chlorine, hydrogen, and sodium hydroxide solution.  The chlorine and hydrogen can be combined to make HCl.  Some chlorine and sodium hydroxide can be used for bleach, if you're interested in clean clothes and disinfected water.

The mgSi is crushed and fed into the top of a reactor, and HCl is fed in near the bottom. They are reacted at moderate temps, a few hundred degrees C, to form a mixture of silicon tetrachloride (SiCl4 "STC") and trichlorosilane (SiHCl3 "TCS").  The TCS is the desired product, but STC can be converted to TCS or fed back into the reactor or used for other things, like fumed silica.  The TCS is desired due to its easier decomposition in the Siemens process. The reactor can be made of glass, but is commonly just plain steel.  Note that some silicones are made from TCS or STC.

Distillation takes advantage of the fact that most metallic chlorides are solids, and left behind as solid waste from the bottom of the reactor.  Aluminum Chloride is an annoying exception, boiling at 120 deg. C.  TCS boils at a conveniently low temperature (32 deg. C, 90 deg. F), so distillation is easy.  Around 4 stages of distillation are required before the TCS is ready for
the Siemens process.  [14]

_Silicon_Tetrafluoride_

The second method of getting pure silicon starts with a byproduct of phosphoric acid production from phosphate rock.  While there are valid concerns about "peak phosphorous", there will likely be production of phosphoric acid from phosphate rock for a long time, though not on the massive scale at present.  And by recycling the fluorides, we can stretch out depletion.

The phosphoric acid process is simple: mine, crush and concentrate phosphate rock, then dump in some sulphuric acid. [15]  While the main results are phosphoric acid and phosophogypsum, any fluorine in the source rock is converted to hydrofluoric acid (HF), which reacts with any silica impurities to form hexafluorosilicic acid. [16]  This is commonly used in water fluoridation today.

Hexafluorosilicic acid (H2SiF6) is separated from the raw phosphoric acid, then can be decomposed into silicon tetrafluoride (SiF4), typically by dehydration with concentrated sulphuric acid. [18]  SiF4 is a gas that boils at -86 C (-123 F), so can be distilled to purify it.  As long as kept dry, it is non-corrosive (it hydrolyzes to HF and silica)..

Then one makes a metal hydride.  Sodium hydride (NaH) works good, others that have been used are calcium hydride and lithium aluminum hydride and sodium aluminum hydride.  The sodium can be derived from the electrolysis of common salt (NaCl).  The hydrogen from electrolysis of water or from a chloralkali cell. The hydrogen may be simply bubbled through or blown across the top of molten sodium (melting point 98 C, 208 F), though a crust of NaH forms.  A better method involves a "bomb" (sealed, high pressure reaction vessel) at 250 deg. C and 500 psi of hydrogen. [19]

One then reacts the metal hydride with SiF4 to get silane (SiH4).  This is done my mixing the hydride as a slurry in an inert chemical, typically an ether such as diphenyl ether (boiling point 258 deg. C). [20]  This reaction takes place at around 250 deg. C, and near atmospheric pressure,  as long as a high-boiling liquid is used.  Diphenyl ether is just two benzenes joined by an oxygen, easily synthesized from phenol (a benzene ring with an -OH radical).  Currently, phenol is derived from petroleum, specifically benzene and propylene.  But it could be made from toluene, which can be extracted from certain pine tree oils.

One will also need an inert gas to purge the system before generating any silane, since silane is pyrophoric.  This is where steam engine technology of pistons in cylinders comes in.  By compressing and cooling air in stages,  it can be liquified, as was done in 1883. [21]  Then the liquid air is distilled to separate the components.  For use in silicon manufacturing the argon component of air is the most interesting because it is a true inert gas.  But oxygen and nitrogen have other industrial uses, so they may be valuable byproducts.  One use for cold byproduct gases would be to freeze impurities out of the silane as part of purification of it.

The resultant metal fluoride salt may have use, for example, SunEdison (formerly MEMC) uses sodium aluminum hydride that makes sodium aluminum fluoride, a.k.a. artificial cryolite, which is used in aluminum reduction.  Or the sodium fluoride could be electrolyzed, and the fluorine used to make HF for PV or semiconductor fab use, and the sodium recycled for hydride.

_Siemens_Process_

So now one has either pure TCS or pure silane, one has to reduce them to pure polysilicon.   While there is another process using a fluidized bed reactor to make polysilicon beads, it's somewhat complex, so I will focus on the tried and true, which is the Siemens process, developed in 1945.

Seed rods of silicon, about 1 centimeter (half inch) in thickness and several meters long,
are put in a sealed chamber.  They are held at the (cooler) floor (which prevents deposition at the holders), so a small bar of silicon is used to bridge pairs of the seed rods (it just rests on top in notches).  The chamber is evaculated, the rods are heated with radiant heat until they become conductive enough, then electrical current is used to heat them.  The chamber walls are then cooled to prevent deposition on them.  Once the rods are hot enough (around 1000 deg. C), the feed gases flow in and react on the hot substrate leaving solid Si. [22]  When reducing TCS or STC, one must also introduce hydrogen.

        SiHCl3 + 2 H2 --heat--> Si(s) + H2(g) + 3 HCl(g)
        SiCl4 + 2 H2 --heat--> Si(s) + 4 HCl(g)

Silane just decomposes, with no corrosive offgases.

        SiH4 --heat--> Si(s) + 2 H2(g)

Often an excess of hydrogen is used, even with silane, to suppress gas phase nucleation that would cause a lot of dust in the process.

One the seed rods have grown to a decent size (around 100 - 150 mm, 4 - 6 inches), which takes a few days, the flow of gas is shut down and the reactor cooled.  Once cool, it is opened and the polysilicon rods removed.

_Crystallization_

To be useful for electronics, the polysilicon must be crystallized into a single crystal of known orientation so the semiconductor devices are repeatable and consistent.  For PV, the most efficient devices are also mono-crystalline, although decent devices can be made with multicrystalline material.

Also, crystallization time is typically when the silicon is doped, that is intentionally contaminated with a trace of an element that will donate or accept electrons, to make an n-type or p-type semiconductor. [23]

There are 3 major crystallization methods in common use:
Czochralski (CZ) makes mono-crystalline cylindrical boules
Float zone (FZ) makes mono-crystalline cylindrical boules
Directional Solidification (DS) also called Heat Exchange Method (HEM)
                makes multi-crystalline cuboid ingots.

CZ requires pure silica crucibles, in which a charge of broken up chunks of polysilicon rods is melted, along with a small amount of dopant.  Then a seed crystal is dipped into the melt and slowly withdraw.  Both the crucible and seed need to rotate about the vertical axis.  CZ requires some complex graphic heater parts, and inevitable contaminates the ingot (or boule) with oxygen from the silica crucible. The process was invented in 1916, so it's not exactly high-tech. It will take about 2 days to grow a crystal. [24]

FZ does not need any crucible, but it does need a nice, straight crack-free polysilicon rod, which means slower (and less economic) growth in the Siemens reactor.  It requires RF heating for any practical size of boule, so this could be a problem.  The frequencies need to be between 2 and 5 MHz, with a power of at least 50 kW.  So one is looking at a large power tetrode vacuum tube, like were developed in the 1940s, unless one can get a solid state amp (see Part 3.)  Resistance heaters would block the view of the molten zone, critical for operator control.  Optical heaters would work, but without the electromagnetic levitation/repulsion of the molten silicon, the boules could only be about an inch in diameter. Early transistors were made on such small wafers, and so were the 1st silicon PV cells, but they were far from economic.

Float zone typically uses gas phase doping, so one would have to come up with a source for borane (BH3) and/or phosphine (PH3).  Borane should be easy, boron oxides from natural borates are dissolved in HF, then the resulting boron trifluoide is reacted with sodium hydride, all familiar.  Probably the best source of phosphine would be to make elemental white phosphorus (in an EAF) and react that with sodium hydroxide.  Both gases are nasty and toxic, but fortunately dopants are only needed in part-per-million quantities. [25]

The tradeoffs between CZ and FZ are summarized in ref. 26.  Both of them require argon gas flows to purge the system. In a resource constrained world, the small size of the FZ boules may not matter so much, but the RF issue might.  Currently FZ produces only a tiny portion of all the monocrystalline silicon used in the world.  I'm thinking in a resource constrained world it may still be easy enough to create silica crucibles (pure sand fused with oxy-hydrogen
torches) that CZ will continue to dominate. [26]

Note that float zone is often used to make the Siemens seed rods, as one Siemens polysilicon rod can be drawn out into many seed rods.

DS can only (really) make multi-crystalline ingots, so is only applicable to PV.  But they're square, which aids packing them into a module and saves on encapsulent materials.  The process is to melt some broken chunks of polysilicon in a square crucible, and then cool it from the bottom.  Like water, silicon expands when it freezes, so a crucible cooled from the top would break.  These crucibles are only usable one time, and have to be broken off the ingot.  The sides and bottom of the ingot get contaminated by the crucible, so are cut off.  But current melt sizes are up to 650 kg of polysilicon, which is several times larger than a CZ boule, so there are some economic advantages. [27]

The downside to DS is the lower efficiency of the PV, and the large square use-once silica crucibles required. At small scale, one might want the flexibility of a CZ grower, but at larger scale for a dedicated PV manufactury that would be irrelevant.

_Shaping_and_Sawing_

So now that we have a silicon ingot or boule, we have to cut the thing up into wafers, so they can be further processed.  First is a rough trim.  For mono-crystalline boules, the "tops" and "tails" are cut off, and these can often be recycled in the next melt.  Multi-crystalline ingots often trim the bottom off, but sometimes that step waits until "bricking".  The usual method of cutting at this stage is a diamond impregnated circular or band saw.  But industrial diamonds might be scarce in a future world.  Fortunately, silicon carbide will work to cut silicon, though it will be somewhat slower.

Mono-crystalline boules need to be rounded off perfectly, so they are chucked on a lathe-like machine, and ground smooth.  If they are to be used for semiconductor applications where orientation matters, X-ray diffraction will be used to ascertain the crystal orientation, and a flat or groove will be ground to indicate the alignment.  N.b. X-rays were discovered in 1895, though they had been inadvertently being generated since 1875 in Crookes tubes.  A high voltage power supply above 10 keV with a few mA output is fed to a vacuum tube with a heavy metal target.  A Geiger counter tube or fluorescent screen can serve as the detector of the diffraction pattern that indicates the crystal planes.

Semiconductor boules will be sent to wafering, but PV boules/ingots need to be "bricked" - cut into square (or nearly square for the round boules) sided blocks. This might be done with the same saws as the trim stage, or wire saws might be used. [28]

The bricks/boules are cleaned, then mounted to a sacrificial bar with glue, so the wire can cut completely through the silicon, yet the wafers will stay in place.

Wire saws are an invention from the PV community that has propagated "back" to the electronics community.  The basic concept is a spool of wire that is guided by pulleys where cuts need to be made.  This can be used for bricking, or with pulleys with many parallel grooves, for wafering a brick all at once.  Just before the wire meets the silicon, a slurry of silicon carbide grit is dispensed onto the wire.  The slurry base is typically polyethylene glycol, so in a simpler world, this might have to be a plant based oil.  Currently a number of diamond impregnated wires have come to market, which only require water as a coolant.  But the production of large quantities of tiny industrial diamonds might be difficult in the future, as will the mining of natural diamonds. [29]

The steel wire is fairly fine, these days 110 - 140 um (0.0043 - 0.0055 inch), so this might be an issue, though these sizes are larger than the smallest size in the British Standard Gauge wire sizes, which dates from 1883.  Other consumables for wire saws are the sacrificial bars the bricks are glued to and the glue that holds them.  Glass is a commonly used bar that could be available in a low-tech future.  Glues are more of an issue.  Currently epoxy is often used, but a thick pitch will work.

After cutting, the bars of wafers are removed, the glue dissolved and the wafers demounted and cleaned.  Currently petroleum solvents are often used, but they can be replaced by plant based materials, ethanol, and/or a soak in some hot sodium hydroxide (lye) to get organics.  Further cleaning will be covered in the relevant sections for PV or electronics.

Semiconductor wafers would undergo an edge grinding process to round the edges over, which lessens breakage while processing, but is mostly done so that spin coating works well.  Then they would undergo some amount of polishing, though not to the fine level of wafers for today's advanced processes. Fine silicon carbide and alumina would suffice.

Discussion

Currently, mgSi is a couple of dollars per kg, and polysilicon prices are near cost at about $20/kg.  For PV, polysilicon usage is now around 6 down to 5 grams per "Watt peak" (Wp), meaning the amount of solar cell that will produce a watt of power at the standard illumination of the standard spectra ("AM1.5").  Assuming thicker wire, and less optimization,
this might go back up to 10 grams/Wp, so a megawatt of PV would require 10 tonnes of polysilicon.  Silica's molecular weight is 60 gm/mol, silicon is 28, carbon is 12.  So 10 metric tons of polysilicon require 21.5 tonnes of silica, and roughly 8.6 tonnes of charcoal, plus a few tonnes of wood chips.  In 2012, there were 3,313 MWp of PV installed in the U.S., which was about 11% of the global installations. [31]  (This is about the max output of 3 large coal or nuclear plants, though the "capacity factor" - percent of power actually produced - would only be in the 20% range for the U.S., due to sunlight only being around during the day, clouds, latitude, dust soiling the modules, etc.)

To just maintain this 3 GWp install rate (domestically) would be a tremendous undertaking without fossil fuels.  Even at 5 gm/Wp, 3 GWp via the EAF process would require 32,250 tonnes of silica and 12,900 tonnes of charcoal per year.  Call this 50,000 tonnes/year to be moved, give it 250 work days, and one gets 200 tonnes per day.  That is nothing for railcars (about 2 cars), or 10 trucks, but animal drawn wagons typically carry only a ton or two.  By no means is this impossible, but it would likely have to be distributed.  The charcoal production is a small fraction of current world production, but then again, without fossil fuels, that demand might well rise. [32]  The current cost of the charcoal would be $860 to $3,640 for 1 MWp, but those seem to be third world prices or automated first world.  Even at the high end, that's still a fraction of a cent per Wp, but again the future is uncertain.  How efficient will the wood choppers be without gasoline powered chainsaws?  How many will be required?  A small production of a few MWp per years would be very manageable, some 10's or 100's of tonnes per year of raw materials.  But the smaller inputs of acids and the like would be a larger cost burden due to lack of economies of scale and complementarities with other industries.

Currently PV modules go for about $1/Wp, roughly $0.75 for the wafer and cell processing.  Wafers are in the 30 cent/Wp range these days.

Silicon semiconductor wafers are typically an irrelevant part of the cost of semiconductor devices.

Summary

Going from sand to a silicon wafer has been described in detail.  As long as economic justification can be made, and the equipment to build the described equipment is available - drill presses, lathes, and the like, and the requisite electrical power, mineral/natural resources obtained, then I think silicon wafers are within the realm of feasible.  Not as an individual crafts-person's shop, but as a group of something like 20 workers at the smallest scale on up to massive factories.  However, without more information about the economics of any particular "post-industrial" situation, it is hard to say yea or nay.


references
=========
[1] http://minerals.usgs.gov/minerals/pubs/commodity/silicon/mcs-2013-simet.pdf
http://minerals.usgs.gov/minerals/pubs/commodity/silicon/myb1-2011-simet.pdf
yearly updates and historical versions at
http://minerals.usgs.gov/minerals/pubs/commodity/silicon/

[2] http://www.decision.eu/doc/brochures/exec_wei_current.pdf

[3] http://en.wikipedia.org/wiki/Future_of_the_Earth

[4] http://www.greentechmedia.com/research/report/polysilicon-2012-2016
http://www.pv-tech.org/friday_focus/what_next_after_the_polysilicon_apocalypse_of_2012

[5] http://www.pv-magazine.com/news/details/beitrag/polysilicon-industry-to-resume-growth-in-2013_100009775
http://www.bernreuter.com/en/news/press-releases/global-polysilicon-market-report-2013-forecast.html

[6] http://en.wikipedia.org/wiki/Electric_arc_furnace

[7] http://en.wikipedia.org/wiki/Calcium_carbide

[8] http://en.wikipedia.org/wiki/Carbide_lamp

[9] http://en.wikipedia.org/wiki/Oxyacetylene#Acetylene

[10] http://en.wikipedia.org/wiki/Silicon#Metallurgical_grade
pics and more info at:
http://siliconsultant.com/SISiFeed.htm

[11] http://en.wikipedia.org/wiki/Hydrochloric_acid#History

[12] http://en.wikipedia.org/wiki/Lead_chamber_process
http://en.wikipedia.org/wiki/Sulfuric_acid#History
A nice history of sulphuric acid.
http://pubs.acs.org/subscribe/journals/tcaw/10/i09/html/09chemch.html

[13] http://www.flickr.com/photos/pembina/3791986017/
http://folc.ca/sulphur_storage/sulphur_blocks_4.htm

[14] Advanced Silicon Materials for Photovoltaic Applications,
Sergio Pizzini ed. John Wiley & Sons, 2012

[15] http://en.wikipedia.org/wiki/Phosphoric_acid#Wet

[16] http://en.wikipedia.org/wiki/Hexafluorosilicic_acid

[17] http://en.wikipedia.org/wiki/Chloralkali_process

[18] see U.S. Patent 4,470,959 from 1983, or

[19] U.S. Patent 1,958,012 from 1931
Inventor is Anthony Moultric Muckenfuss.

[20] U.S. patent 4,374,111

[21] http://en.wikipedia.org/wiki/Liquid_air

[22] a picture of a Siemens reactor after growth in fig. 3 at:
http://www.photovoltaic-production.com/3970/precisely-operated/

picture of hot rods just starting at:
http://guntherportfolio.com/2010/05/bernreuter-researchs-the-whos-who-of-solar-silicon-production/

more info:
http://www.tf.uni-kiel.de/matwis/amat/semitech_en/kap_4/backbone/r4_1_2.html

[23] http://www.pveducation.org/pvcdrom/pn-junction/doping

[24] http://www.pveducation.org/pvcdrom/manufacturing/czochralski-silicon

[25] http://www.pveducation.org/pvcdrom/manufacturing/float-zone-silicon

[26] http://siliconsultant.com/SIcrysgr.htm

[27] http://www.pveducation.org/pvcdrom/manufacturing/growing-ingots

[28] http://www.pveducation.org/pvcdrom/manufacturing/sawing

[29] http://www.pveducation.org/pvcdrom/manufacturing/wafter-slicing
http://www.appliedmaterials.com/sites/default/files/Advanced-Wire-Sawing-Technology-Whitepaper_en.pdf

[30] http://www.netl.doe.gov/publications/factsheets/rd/r&d112.pdf

[31] http://www.seia.org/research-resources/us-solar-market-insight-2012-year-review

[32] http://www.biochar-international.org/images/Stefan_Czernik.pdf

1 comment:

  1. Excellent work. Your thoroughness is a real gift.

    It truly is hard to know what the far future will bring down the long descent, and whether or not it will be possible for complex projects like this to remain resourced, organized and capitalized.

    It seems like it would require such coordination between so many sectors of society - financial, manufacturing, educational and engineering. This may seem like no big deal for us in today's world, but as you acknowledge, it's hard to know how the economics will play out.

    And more than just that, even though one could argue that the economics should work out to produce PV cells on a smaller scale and more regional basis in the far future when cheap energy is a thing of the past, what I'm asking is whether or not all of the different sectors and their coordination will still be available in that far future to work together to meet that economic opportunity.
    I don't pretend to have an answer.
    I sure would like it to work out that way, though.

    I think I have another line of thought on future technologies that you'd find interesting:
    http://sunlitsynergy.blogspot.com/

    Take care

    ReplyDelete