Cooking utensils. A typical alloy would be 3003. Actually, we are more likely to be recycling siding into cooking utensils (for the military), than the other way around. Teddy Roosevelt had an aluminum canteen when he led the charge up San Juan Hill.

Grantville Aluminum Reserves How much aluminum was trapped by the Ring of Fire? I can estimate this by a simple top-down approach. During the period 1978-1988, the per capita aluminum consumption in the USA was 25-30kilograms per year (Kirk-Othmer, 2:207). Let's go with the higher figure.

In 2004, about 20% of North American aluminum use was in packaging (cans and foils) (Aluminum Association). Packaging is going to make a relatively small contribution to our aluminum reserves. People don't usually buy a year's supply of soda or foil at one time. In fact, I would guess that the average person had only the equivalent of a six pack of soda cans, and a roll or two of foil, when the Ring of Fire occurred. You would need thirty-two all-aluminum cans to provide just one pound of metal. The can-and-foil reserve is so small I am going to ignore it, for now.

The other 80% of aluminum products include building materials, cars, and wiring. The period from when the product is sold to when it's discarded or recycled is called the recovery cycle, and has been estimated as being ten years for transportation (especially cars) and consumer durables, twenty for producer durables (machinery), and up to fifty years for cabling and construction. (Brubaker, 76; Choate, 61) A United Nations agency has used a weighted average life for aluminum products of twenty years (Id.), but I am going to assume a more conservative average recovery cycle for aluminum products, other than cans and foils, of ten years.

So that means that the per capita reserve of aluminum is ten times the 80% of 30 kilograms, or 240 kilograms. Eric and Virginia agree that there were about 3,500 people trapped by the Ring of Fire, so that's a total of 840,000 kilograms (about 925 American tons).

We know that's an overestimate. After all, we are crediting every inhabitant of Grantville with his or her share of aluminum use in locomotives, ships and aircraft, and they aren't actually present. (On the other hand, if aluminum is used in coal mining equipment, Grantville has more of that than does, say, Washington D.C.) But to play it safe, let's knock the estimate down to 800 American tons.

Does that still seem like a lot? In 2000, the average aluminum use was 100 kilograms per car in Western Europe. (Key to Metals) American cars are larger, but probably make less proportionate use of aluminum. If there were just 1,000 cars in Grantville (one for every 3.5 residents), that alone would add up to 100,000 kilograms (about 110 American tons).

Of course, much of the Grantville aluminum is locked up in products which people want to keep.

Recycling Technology and Economics NAP says that if you want to use cans (3104 sides, 5182 ends) to make fresh 3104, that is possible, as the excess magnesium (from 5182) can be removed by various treatments.

On the other hand, suppose your target alloy was 7020 (used in armored vehicles and military bridges) or 7075 (used in airframes)(azom.com)? The manganese content of 3104 deliberately exceeds the standard tolerances for 7020 and 7075. And it may (because of its higher tolerances) have too much silicon or iron.

Reclamation technology is a big subject, which I will just touch upon here. Scrap is sorted based on what impurities it is likely to contain. It is then processed in a variety of ways. The first stage is likely to be mechanical, reducing the scrap to powder. Next comes "cleaning." In pyrometallurgical cleaning, the organic contaminants are removed by heating the scrap to a temperature high enough to vaporize the impurity, while staying below the melting point of aluminum.

Then, to separate aluminum from iron (and other high-melting elements), you raise the temperature enough to melt just the aluminum, and it trickles down from the furnace ("sweating").

In hydrometallurgical cleaning, the scrap is washed with water, thereby removing water-soluble components ("leaching"). After drying, the powder is passed through a magnetic separator, which removes some of the iron. If the material is placed into a viscous medium, you can separate low-density aluminum from high-density metals such as copper and iron. Of course, magnesium, another light metal, is going to keep the aluminum company.

After these cleaning processes, the scrap is ready for smelting. Scrap can be "demagged" by smelting the scrap in the presence of a flux, and a chlorinating or fluorinating agent (liquid chlorine, aluminum chloride, aluminum fluoride). The agent reacts with the magnesium to produce a magnesium chloride or fluoride which rises to the surface and is trapped by the flux. There, it is skimmed off. All of the common "demagging" agents are nasty stuff, by the way. (EPA) * * *

Assuming that there is no impurity that must be removed, and no new alloying element to be added, the energy cost of recycling aluminum is very low. The aluminum, except at the surface, is already in reduced (metallic) form, so all you need to do (after removing, if possible, non-aluminum components) is to melt the aluminum, which has a melting point of 669 deg. C.

The energy requirement is only 0.3-0.4 kWh/kg (Choate, 63), about 3% that for smelting aluminum.

Adding alloying elements to refined scrap isn't much different than adding it during primary production.

In our time line, primary aluminum usually sells for a slight premium over secondary (recycled) aluminum.

Gorg Huff has pointed out to me that products made from recycled aluminum may receive a "bonus"

valuation in the 1632 universe because they were brought into that universe by a miracle, the "Ring of Fire." That would apply especially to, say, a cross or other religious artifact made of recycled aluminum, perhaps complete with a "certificate of authenticity."

Aluminum Substitutes Aluminum has many desirable characteristics. However, for each of these, there is an alternative material, used in the seventeenth century, or obtainable with uptime assistance, which it is seeking to compete with: ? lightness: wood and, eventually, plastics ? electrical and heat conductivity: copper, silver, gold ? corrosion resistance: lead, tin, copper, silver, gold, and, eventually, nickel and galvanized steel ? reflectivity: silver The lack of aluminum would be felt, most acutely, by the aircraft industry. In 1952, military aircraft were about 75% aluminum alloys. (Hultgren, 291) Aircraft can still be built-as evidenced by theLas Vegas Belle and theGustav- they will just be heavier, and thus have to sacrifice payload, or range, or speed.

Why Aluminum?

Since substitutes exist, we have to ask, "Why aluminum?"

In World War Two, aluminum was the "metal of war." World War One didn't see much use of aluminum in warcraft, but it did stimulate research on aluminum alloys. (Wallace, 46-7) However, we aren't recreating twentieth-century military technology. Rather, Grantville's plan for survival is to "gear down": "Use our modern technology while it lasts, to build a nineteenth-century industrial base." (1632, Chapter 11) Our plans for military hardware likewise have an American Civil War flavor-rifled muskets using Minie balls, simple breechloaders, bayonets, ironclads, Requa battery guns.

(1632Chapter 47; Weber, "In the Navy,"Ring Of Fire ; Hollar, et al., "Flint's Lock,"Grantville Gazette, Volume 3; Hollar, et al., "How to build a machine gun . . ."Grantville Gazette, Volume 4) Aluminum will become a military necessity only when our enemies have "geared up" to the point that we need twentieth century military technology-especially massed airpower-to maintain battlefield superiority. But in 1633, even the most enlightened of our enemies (Turenne) is thinking in terms of percussion caps and rifled muskets (1633, Chapter 21), not machine guns and tanks.

Hence, the aluminum industry is not going to be another Manhattan Project. It will need to justify its creation to hard-nosed financiers, who are more interested in return-on-investment and risk-of-loss than in the wonders left behind by the up-timers.

That means that, within a reasonable time after start up, the new aluminum company will need to be able to sell aluminum products at a profit.

The ability to profitably make an aluminum product is going to depend on the price of aluminum, the cost of manufacture, and the utility of the product in Grantville and elsewhere.

Price. In 1884, world aluminum production was about 3.6 metric tons, about one-thousandth the level of silver production. However, the demand for aluminum was also small, and so it sold at the same price by weight as silver. (Binczewski) Historically, the industry took off once the price of aluminum dropped to about double that of copper. It is at that price that aluminum electrical wire is cheaper than its copper conductor equivalent. Between 1945 and 1950, aluminum became cheaper than copper even on a "by weight" basis. (Brubaker, 61) There are some products for which aluminum is initially more expensive, but provides lower operating costs. For outdoor structures, aluminum's corrosion resistance means less repainting. (Those sidings we recycle may return.) In the transportation industry, replacing steel with aluminum results in less weight and thus in lower fuel bills.

Cost of Manufacture. For those writing stories set in the 1632 universe, the issue is whether it isplausible that aluminum could be economically produced. In this author's opinion, the manufacturing cost will not be much more than it was in our timeline in the 1890-1920 period. Those who want a more detailed cost analysis may consult the Appendix.

Utility. One advantage that aluminum entrepreneurs have in 1632 which their old time line counterparts lacked is that they are already familiar with many uses for aluminum. Heroult didn't commercially exploit his discovery because he heeded the advice of an "expert," who told him that "aluminum was a metal of restricted usefulness; at most it might be used for opera-glasses. . . ." (Wallace, 512) It is convenient to classify the aluminum products into two categories, low-tech and high-tech.

Essentially, the low-tech products are those which could be sold anywhere in Europe, as soon as they can be made, because they don't require an improved technological base to make or appreciate.

The low-tech category has a high-end (jewelry), and a low-end (what the industry calls the "pots and pans" market, but which would include camping and hunting equipment). The boundary isn't necessarily a sharp one; Napoleon III had aluminum utensils and a ceremonial aluminum breastplate.

The problem is really one of marketing: how do you make the nobility pay through the nose for the privilege of owning an aluminum piece and still exploit the mass market? One approach is to start by selling a small number of aluminum articles at high prices, then, when you have exhausted the upper class market, lower your prices and go after the bourgeoisie. Another is to find some way of distinguishing the high-end product; combine it with gold or gems, or anodize it distinctively, or have a craftsman "sign it."

Other low-tech aluminum products include mirrors, siding, cans, and, perhaps, carriage frames.

Foil is something of a special case. There will certainly be a market for means of retarding spoilage of food. However, to make foil you need a rolling mill.

The high-tech markets will open up first in the vicinity of Grantville, then widen as the necessary background technology diffuses outward. If historical patterns are followed, the first of these markets will be for treating steel, then electrical wiring, then self propelled land vehicles, and finally aircraft. While nominally high tech, the steel industry market was the principal one in the United States in 1889-1893.

(Wallace, 10) It is also possible to divide the products, independently, into those which are simply aluminumized versions of seventeenth century wares, and those which are altogether new products. In the former case, the goal of the industry will be to point out the special advantages which aluminum confers on an otherwise familiar product. In the latter case, they have to create demand for something new.

Evolution of the New Aluminum Industry In Karen Bergstralh's story "One Man's Junk" (Grantville Gazette,Volume 4), Herr Glauber demonstrates a fine appreciation of the post-Ring of Fire role of aluminum. In a shed, he and his colleagues find an aluminum lawn chair. As a blacksmith, he is quite aware of the properties of the material: it is light, and it doesn't rust, but it is soft. "Unfortunately, it will be years before they can make more of it. Still, that lack makes what remains all the more valuable."

Clearly, the initial activity is going to be recycling aluminum articles. In the case of Glauber's find, thatdidn't even require processing the aluminum: "that should clean up and with a nice new leather seat it will fetch a fancy price in Jena or perhaps Amsterdam."

At the same time, a lot is going to be going on in other industries. We will be coaxing Venetian and Thuringian glassmakers to make chemically resistant borosilicate glass, importing and refining Japanese zinc, and producing a variety of industrial chemicals. We will be mining coal, prospecting and drilling for oil, and scaling up steel production.

This activity is both good and bad for the aluminum industry. On the one hand, it will increase the demand for aluminum products. For example, if we have gasoline, and decent roads, motor vehicles will be returned to operation, and there will be a need for replacement aluminum parts.

The post-Ring of Fire industrial ferment will also create the infrastructure to support the aluminum industry. For example, hydrogen fluoride (which is stored in carbon steel, or other HF-resistant metals, not glass) will come in very handy. Steel-hulled ships would make it safer to fetch Jamaican bauxite or Greenland cryolite. It would be nice to have pure zinc, magnesium, manganese and silicon for alloying use, too.

On the other hand, these new or improved industries are going to be competing with the aluminum industry for electricity, which is the very lifeblood of the Halt-Heroult process.

So, I see an uncertain "window of opportunity" for early development of the aluminum industry. It starts when we have in place the infrastructure, and ends when it becomes too expensive to buy electricity from the Grantville power plant. If I had to make a guess, this window will run (please don't shoot me) from ROF+5 to ROF+10.

The single most critical issue determining whether this early development occurs is probably going to be whether the up-timers figure out how to make cryolite synthetically. That at least eliminates the need for a Greenland expedition. And they do know how to make hydrogen fluoride and carbon steel.

If everything comes together, we are probably talking about production, ultimately, on the scale of Hall's 1895 New Kensington plant-up to 2,000 pounds a day. At 10 kilowatt hours a pound, that's an energy draw of 20 megawatt hours each day. Bear in mind that it took Hall five years (1888-1893) to go from fifty pounds daily to 1,000 pounds, and then two more to reach 2,000 pounds. (Beck) And we can't work out the inevitable "bugs" in the electrolytic process until we have on hand the raw materials: bauxite (to make alumina), cryolite, and suitable carbon.

Moreover, the first alloys we make are likely to be the easier ones, where the alloying element, if any, is copper or tin.

If this first window of opportunity is missed, then the next one will open when other power plants come on line. That probably means ROF+15 or later.

It is thus uncertain whether newly manufactured aluminum will be the "wonder metal" of the 1630's, or just a will o' the wisp, something pursued but never captured.

APPENDIX: Manufacturing Cost Analysis It is extraordinarily difficult to determine how much it would cost to manufacture aluminum in the 1632 universe. Suppose we know the capital and operating costs for bauxite mining, alumina production and aluminum smelting in the twentieth century. If we try to translate the amount into seventeenth century terms, we run into the usual problems of determining the relative purchasing power of, say, the 2000 American dollar and the 1632 shilling or guilder. The economies are so different. Still, the usual rule of thumb is that one 1632 Dutch guilder is worth $42 in 2000 currency, and that one English shilling is worth half that.

Inflation also affects, to a lesser degree, the purchasing power of the cited nineteenth and twentieth century prices. Aluminum sold for about $700 per metric ton in 1900, but that is worth more like $14,800 in 2000 dollars. (USGS) Or about 700 shillings in 1632. Inflation from 1963 to 2000 was about 5.6-fold.

We can instead use a "bottom-up" approach: try to determine the labor and material requirements, and cost those out using seventeenth century sources. However, that is harder than it seems. Even if the raw material is one which is commercially available in the seventeenth century-soda, for example-its availability and cost is also going to be altered by the influx of uptime technology. And how does one cost out cryolite, or electricity?

Capital costs are another problem. Calculating the contribution of capital costs to the cost of production requires making assumptions about the loan term, the interest rate, and the ratio of production to investment. Ugh. Rebecca does offer Gustavus Adolphus loans at four percent annual interest, for economic projects, but private parties may find the rate "quite a bit higher." (1632, Chap. 47) Our data are mostly based on the perfected Hall-Heroult process, as practiced recently in industrial countries. The up-timers of Grantville will undoubtedly recreate the process in a cruder, less efficient form, and they will have much less in the way of infrastructure to support them. The best we can do is to draw our data, where possible, from the early days of the industry, or from Third World countries.

Bauxite mining Bauxite, generally speaking, is inexpensive. In 2000, the price of Bauxite imported into the United States was $23 per metric ton. That includes both mining and shipping costs, as well as the seller's profit. In 1963, the cost of mining and drying bauxite was estimated as being 1.75-4.50/ton, and its selling price, at the port of shipment, at $7.23-13.82/ton. (Brubaker 149) The main difference between bauxite mining today and bauxite mining in the 1632 universe is that the latter will not have the advantage of power equipment. However, this advantage ought not be exaggerated. In 1900 Arkansas, bauxite was mined manually, and transported by wagon to the railroad station at Bryant.

In 1913, French bauxite was imported into the United States at a price of $3.94 per short ton. In 1915, the price of Arkansas bauxite, ready to be shipped by rail, was $4.70-5.50 a ton, whereas in 1902 it was $2.50-3.00. (ICC Reports, 37:142-3) I suspect that these prices were for bauxite mined by hand.

Multiply by 15-20, and you get the equivalent 2000 prices.

Bauxite mining, at least in France, Arkansas, Jamaica, Guyana and Suriname, will be surface mining. I would therefore expect it to be somewhat cheaper, per ton of ore, than the underground mining of the seventeenth century. After all, you are just digging an open pit, not a tunnel. You don't need scaffolding to keep the tunnel from collapsing, and the tunnel dimensions don't limit the number of workers who can be extracting ore simultaneously. You don't need to pump air in, or water out. Surface mining was definitely more labor efficient. In the case of mid-eighteenth century iron mining, "mine holes forty to fifty deep required eight miners to produce the same daily tonnage as six had done from a surface trench." (Kennedy) I realize that the cost of labor in the New World will be higher than that in Europe. However, it is mistaken to think that the only way to obtain enough miners is to import slaves from Africa. The majority of the miners in Mexico and Peru were free Indians who were paid wages. (Bakewell,Mines of Silver and Gold in the Americas) It is likely that the cost of shipping will be as high as the cost of mining. In 1963, the cost of shipping bauxite from Surinam to Holland (4,600 miles) by steamship was $6.50/metric ton, or fourteen cents per ton-mile. (Brubaker 158) That probably added 50-100% to the price at the source. Our early shipments will, at best, be by sailing ship.

It is difficult to predict what the price of bauxite will be, but it seems safe to assume that the priority will be to find deposits which are either near navigable rivers in Europe, or near the coast overseas. The cost of overland movement can be three to ten times that of the same movement by sea.

Alumina Production In 2000, the price of metallurgical grade alumina was $226 per metric ton, whereas in 1963 it was about $60 (USGS).

Usually, the bauxite ore is processed near the mine itself, thus reducing shipping costs. Bauxite and alumina move at about the same cost per ton-mile, but if you process first and then ship alumina, you are moving only half the weight (or less).

In 1963, producing one metric ton of alumina required 2.1 tons of trihydrate bauxite (gibbsite), 80 kilograms caustic (sodium hydroxide), two tons of steam, 200 kilowatt hours of power, 130 liters of fuel oil, 3 man hours of labor, and $3.00 in equipment maintenance. If the ore was the monohydrate (boehmite, diaspore), the requirements were 2.5 tons bauxite, 140 kilograms caustic, 2.4 tons steam, 275 kilowatt hours power, 130 liters fuel oil, 4 man hours labor, and $4.00 maintenance. If the ore is high in silica, more caustic is needed, and alumina recovery is lower. (Brubaker, 88-9) In 1963, the capital cost per metric ton of annual bauxite processing capacity was $110-150 for a trihydrate plant and $140-180 for a monohydrate plant (assuming large plants of more than 330,000 ton capacity). A plant of "only" 100,000 ton capacity would cost $50-60/ton more. The recommended minimum size for a plant in an third world country was 30,000 metric tons. (Brubaker) Wallace said (194) that after alumina plants reach a size sufficient to produce 15-18,000 tons alumina per year, there is no further economies to be gained by further increases in scale. He also said that labor was only 10% of the final cost of alumina, and that the cost of producing alumina was primarily attributable to the raw materials (bauxite, soda, coal).

In 1995, alumina production was at a total cost of $194/metric ton, with the breakdown 30% bauxite, 18% energy, 18% labor/overhead, 15% capital, 10% other materials, and 9% "target profit." (Hadley, 10).

Smelting Costs, Generally In 1963, producing a metric ton of aluminum in a pot with prebaked anodes required 17,000 kilowatt hours of electricity, two tons alumina, 500 kilograms carbon, 18 man hours labor, and $20 in operating and maintenance supplies. (Brubaker 96) Cost breakdowns vary somewhat. Hadley (11) says that in 1995, the total smelting cost was $1215/metric ton: 31% alumina, 21% energy, 21% labor, 18% depreciation/overhead, 9% carbon.

Welch reports that electricity, alumina and capital are each about one-quarter, with the last quarter divided between carbon and labor in roughly a 2:1 ratio. Electricity is often quoted as being as high as one-third of the total.

Smelting: Electricity Loren Jones says to assume a payroll budget of $625,000; that corresponds to an average charge of $625 a household, or an effective 5.9 cents per kilowatt hour. At 1999 prices, the coal needed to meet the 1999 residential electrical use would be another $75 per household (see Addendum).

A single aluminum smelter comparable to Hall's 2,000 pound a day plant (1895) would be using 7,300,000 kilowatt hours a year-about 63% of the immediate post-Ring of Fire residential consumption.

The payroll is virtually a fixed cost, whereas the coal is a variable cost. In theory, to attract the aluminum smelting business, the power plant could offer a price of one cent a kilowatt hour (the coal burnt would cost seven-tenths of a cent). As the demand for aluminum increased, and other industries increased their demand for electricity, the power plant could raise the price, perhaps reaching or even exceeding the pre-Ring of Fire standard industrial rate of four cents a kilowatt hour.

Other industries will be demanding electricity, too. Eventually, the price of electricity might reach the point at which it would be advantageous for the aluminum smelting industry to relocate.

A coal-based power plant requires three times as much energy (10,290 Btu/kWh) to deliver the same amount of onsite electricity as does a hydroelectric plant (3,412 BTU/kWh) (Choate 9). Hence, I expect that the industry will eventually move to where there is ample hydroelectric power.

Smelting: Capital Costs In our time line there were economies of scale in building larger smelters, at least up to those which produce 200,000 metric tons per year. The smallest smelters for which I have capital cost figures are those in the 20,000 ton range, costing $1,000-1,300 (assuming use of pre-baked anodes). The capital cost per ton for a 200,000 ton facility is about half that.

Smelting: Labor Costs Labor requirements aren't great, but still they were one of the reasons for the gradual migration to largerpots. Smelter workers will need to monitor and correct the electrolyte content, lower the anodes as they are consumed, make new anodes in the carbon baking plant, break the top crust of cryolite when alumina is added, and siphon off and cast the metal. Obviously, the 1632 universe smelters are going to be much less automated than that which is the norm in twentieth-century industrial countries. Hence, while the 1963 American norm was about 18 man hours (and $65 wages) to produce one metric ton of aluminum, in third world smelters, which are a better model for what we would be building, the likely labor investment is 30-50 man hours/ton. (Brubaker, 93-4, 155)

Smelting: Carbon The 1911 EB says that in theory, two pounds of carbon are consumed for every three pounds of aluminum produced, but in practice, it is a pound for a pound. According to the modern Encyclopedia Britannica, the ratio is one pound carbon for every two pounds aluminum. The theoretical minimum consumption rate is actually one pound carbon for every three pounds aluminum, but the average ratio is 0.45:1.

In 1963, carbon for prebaked anodes cost $55-70 per ton. (Brubaker 91)

Smelting: Cryolite The initial requirement for cryolite is a function of the amount of alumina which can be processed simultaneously. Hall's first pots each held 300-400 pounds (136-181 kg) of cryolite for dissolving the alumina. These pots operated at 1750 amperes. (Beck) I am going to stick my neck out here, and say that the pot size, and more particularly the required initial charge of cryolite, are both proportional to the amperage running though the pot. In Prasad's 216,000 ampere experimental cell, which was intended to represent a modern commercial one, the ratio was 78 kilograms per thousand amperes. I don't know Hall's bath depth, but if he used 85-87% cryolite, and the bath depth is adjusted to yield the right total amount of cryolite, then his ratio is 77 to 103 kilograms per thousand amperes. That's not much of a change considering the amperage was increased more than one hundred fold.

Cryolite is lost as a result of absorption by the carbon lining of the cell, vaporization, and so forth. The consumption data is somewhat contradictory. The low estimate is that it is the equivalent of 2.5-4 kilograms fluoride per metric ton of produced aluminum. (Kirk-Othmer, 2:192; Brubaker, 96) If the fluoride all came from cryolite (it doesn't), that would be 4.6-7.4 kilograms cryolite per aluminum ton.

But elsewhere Brubaker says that consumption of cryolite is 30-100 kilograms per ton (97).

The price of cryolite is also somewhat uncertain, with the quoted values ranging from $130 to $360 per ton. Cost of electrolyte in 1963 is set forth as $16-25 per ton aluminum (Brubaker 154). If the price is assumed to be $200 per ton electrolyte, that implies a usage rate of 80-125 kilograms/ton.

If the reader wishes to estimate post-Ring of Fire production costs by the "bottom-up" method, the cost of alumina can be determined from the last section. For the cost of electricity, I would assume one to four cents per kilowatt hour. Graphite was an article of commerce in the seventeenth century, and there isdata on contemporary laborer's wages.

Bibliography Encyclopedias available in Grantville [EB11-Al] is "Aluminum," [EB11-B] is "Bauxite," [EB11-C] is "Cryolite," all from 1911Encyclopedia Britannica [EA-A] is "Aluminum," and [EA-B] "Bauxite," inEncyclopedia Americana . Also cited: "Hydrofluoric Acid," "Bronze."

[EB-ICP] is "Industries, Chemical Process," [EB-IEP] is "Industries, Extraction and Processing,"

[EB-CE] is "Chemical Elements," [EB-F] is "Fluorspar," all from the modernEncyclopedia Britannica .

Aluminum Industry ALCOA, "Aluminum Smelting,"

http://www.alcoa.com/global/en/about_alcoa/pdf/Smeltingpaper.pdf Aluminum Association, "North America Aluminum Industry - A Quick Review," (Nov. 2005) Ammen, Casting Aluminum (1985) Belov, Iron in Aluminium Alloys (2002) Brubaker, Trends in the World Aluminum Industry (1967) Choate, "U.S. Energy Requirements for Aluminum Production: Historical Perspective, Theoretical Limits and New Opportunities" (USDOE, 2003), http://www.eere.energy.gov/industry/aluminum/pdfs/al_theoretical.pdf Hadley, "Aluminum R&D for Automotive Uses and the Department of Energy's Role,"