The Tellurium Supply Conjecture

The market fundamentals of the technology metal tellurium are an enigma. There is no general agreement whatsoever on the size of the global supply, the rate of production or the location(s) of that production. Yet, the small but consistent uses of tellurium in steel and copper alloying, as well as in military applications, has now been joined by a new demand that threatens to be voracious: the use of tellurium as a critical material for manufacturing high efficiency cadmium telluride thin-film photovoltaic cells and modules.

This situation is ideal for the promotion of both a technology and a commodity bubble based on pure conjecture about the supply of tellurium. Fortunately for us and unfortunately for the bubble-ists we can, I believe, accurately project the maximum possible total production of tellurium in any given time period, even though we do not know the actual production level (supply).

Merriam-Webster’s Collegiate Dictionary (10th edition) defines a conjecture as “a conclusion deduced by surmise or guesswork.” It is amazing to me how many high-technology business models are based on nothing more than conjectures about the security of supply of critical metals. Individual as well as institutional investors, and all of their financial advisors, can learn an object lesson about this class of short-sighted planning, from the story of the expensive development and implementation of some new end-uses for the rare element tellurium, which end-uses should have been easily foreseen to be resource limited dead-ends.

The analysis of a business model or of a specific business plan, involves both subjective and objective determinations of the probability of the success of the model or plan. The evaluation of management skills, for example, is mostly subjective, unless the management team in question has previously brought an identical business model or plan into successful operation; this is very unlikely in high technology because it is usually a new technology, always called an “advanced” technology, that is being proposed. For such technology-based companies, marketing plans are even more subjective because they must base future demand for their products on meeting manufacturing and pricing objectives in the future, which are and must be based on present predictions of what manufacturing costs and selling prices might be in a future world – even when such numbers now change daily with unpredictable trend directions!

The most tenuous and subjective conjecture of all in a technology based business plan, is the guess as to how much people will pay for a particular technology at a future time when other competitive, even more “advanced,” technologies may well have been developed.

So, we must ask whether there are any objective considerations, based on hard (verified or verifiable) numbers, that can be factored into determining the risk of an investment in a technology that will only be brought into mass production in the future? The answer is yes. The probability of success in securing a supply chain for any critical metals involved in the manufacturing of the technology, can be determined quite accurately today. Keep in mind that a critical metal is defined as one without which the technology cannot be manufactured.

Although I will now demonstrate how to evaluate the probability of success of any existing or planned venture for which the rare metal tellurium is critical, I want to first make a general statement about the most common error made by individuals who, in their analysis of a business model or plan, must take into account the security of the supply of any critical metal for any venture.

It is commonly and incorrectly stated by people who do not understand geology, geochemistry, mining or supply chain economics that the percentage of a metal in the earth’s crust, or, even in the universe at large, is related to the ultimate availability of that metal to the human race as a natural resource. This is not only wrong, it is ignorant. I call this rather common reason to dismiss the security of supply problem the earth fundamental argument after a phrase I saw used just last year, in an article in the prestigious peer-reviewed journal Science. It was used to assure the reader that the amount of the rare metal gallium, in this case, in the earth’s crust, was so large as to make practical the idea of using gallium aluminum compounds to decompose water, releasing its hydrogen for use as fuel. This idea was based on a “discovery” by a professor at Purdue University who actually got the university to finance the filing of a patent, proposing this use of gallium aluminum. The professor announced to the world that he would seek financing to start a company to manufacture gallium aluminum alloy, dispense it at fueling stations and recapture the spent alloy and recycle it for re-use.

I wrote an article at the time pointing out that the annual global production of gallium was less than 200 metric tons (t), all of which was in use for the electronics and weapons industries, and that the only source of gallium that even allowed the 200 t p.a. production figure to be achieved and maintained was the production of the base metal aluminum, from which gallium is recovered as a tiny byproduct. The global annual production of aluminum in 2008 was 39 million t, which allowed for the recovery of less than 200 t of gallium. There are no primary mines for gallium.

I hope this brief example and discussion will allow you to understand that the amount of any metal available to the human race is due to its concentration over geologic time into deposits, called ore bodies, of sufficient size, near to or on the land surface of the earth, where power, water and transportation are available and from which the metal ores can be extracted, separated from the rock, refined, and purified by equipment and technologies that are known to work and economical. No two ore bodies are alike and little speculative research goes on in mining engineering. All such research is “to order” and is expensive and time consuming, due to the limited number of skilled researchers and research locations for such endeavors.

I’m now working on what I hope will be a comprehensive book, which I also hope will be useful to investors, that will cover the theme of security of supply and illuminate, for the purpose of eliminating them, such misleading conjectures as the earth fundamental argument. The book’s working title is “The Age of The Technology Metals.” I have recently been named as a Senior Fellow of the Institute for the Analysis of Global Security (IAGS)  and I hope (again) that the IAGS will sponsor my book so it can be published.

Now, let’s take a look at the “tellurium supply conjecture.”

Almost all (90%+) of the tellurium available to the human race exists as a byproduct of copper. The remainder is mostly a byproduct of lead and some small amounts exist as ore forms in some deposits of gold, silver and bismuth. In the case of telluride ores of gold, silver and bismuth, the proportion of tellurium can be large but the substantive amount of tellurium in any of these small deposits is usually uneconomical without the recovery of the gold, silver or bismuth and even with such recovery intended, the deposits may be too small to be economical for any combination of recovered metals.

I am frequently told of a primary tellurium deposit in China, but upon investigation that deposit turned out to be a bismuth telluride ore body, too small to be economical in the face of the costs of development.

The total amount of tellurium that could be recovered from all of the sources above is estimated to be as much as 3,200 t per year by the most optimistic reporters, but the actual amount recovered from all sources is surely no more than 25%-50 % of that total, and it may well be that the total annual maximum possible production is no more than 1,600 t per year as was estimated by the US National Renewable Energy Laboratory (NREL) in a comprehensive 1997 study.

In 2002 in a book by by Ayres, Ayres & Rade entitled The Life Cycle of Copper, its Co-Products and By-Products, published by Kluwer for the World Business Council for Sustainable Development, the following paragraph appears:

4.8.12. Tellurium [Andersson 2000]: Tellurium is the scarcest of all the by-product metals, except for gold . Crustal abundance is 0.005 ppm. It is mainly recovered from copper ores (1.5-3 ppm), where it is considerably enriched. There are two known deposits where tellurium is found at much higher concentrations, one in Mexico (0.2%) and one in China. Tellurium has a significant potential use in thin-film cadmium telluride (CdTe) photovoltaic cells.

Note that the first sentence means “…of all the by-product metals [found in copper], except for gold.” Also the next sentences should read for clarity “Crustal abundance is 0.005 ppm. It is mainly recovered from copper ores [where it is found as a by-product in the range] (1.5-3 ppm), where it is considerably enriched [compared to its crustal abundance].”

If you calculate the maximum amount of tellurium that could be recovered from the 2008 global production of copper using the percentage composition of tellurium in copper as stated above, you only get 16 million t of copper, multiplied by 3 ppm of tellurium per t, equals 48 t. This figure, although low, is still higher than the total production of tellurium produced (outside of the United States) as “discovered” and verified by the United States Geological Survey (USGS) for 2008, which was 38 t from Peru (30 t) and Canada (8 t). The USGS admits that its data do not include US production for reasons of not disclosing information that could give foreign companies a competitive advantage, and it further states that “Australia, Belgium, China, Germany, Kazakhstan, the Philippines and Russia produce refined tellurium, but output is not reported, and available information is inadequate for formulation of reliable production estimates.” Note that including Belgium and Germany in the aforementioned list is (unintentionally) misleading because tellurium production in those countries is based solely on recycling.

I was told last November (2008) by a colleague gathering statistics on global metals production, that she had directly asked Codelco, the Chilean state owned and world’s largest copper producer, how much tellurium it produced in 2007 and was told “less than two tons.” Note that the USGS Commodity Mineral Survey for Selenium and Tellurium for 2007, published in October 2008, does not even include Chile as a source either of selenium or tellurium.

This report further enlarges the possible production of tellurium globally. beyond the total stated in the Ayre’s [2002] book noted above. It says that:

“Total world production of … tellurium has been estimated between … 450 [and] … 500 t/yr…. Based on global copper refinery data (Moats and others, 2007, p. 202-241) the USGS estimates that copper anode slimes could generate … 1,200 t/yr of … tellurium,….”.

At this point, I want to explain why all of the resource reporters addressing the tellurium supply issue use so many “shoulds” and “coulds’ to modify their production total statements.

Tellurium is only efficiently recovered from copper that is refined by the electro-winning process. In this traditional process, “crude” or “blister” copper from a smelter is used as the anode in a bath of sulfuric acid. The copper is plated onto a thin pure copper or stainless steel cathode. In this operation the “impurities” that may be present in the crude copper anodes such as the gold, silver, platinum, arsenic, molybdenum, selenium and tellurium can be made either to dissolve in the acid or “fall out” as “anode slime” or “anode mud,” which is then collected and processed for these trace metals that have been concentrated by this process. To show you the scale achievable, I can tell you that when I was working with the Amax secondary copper smelter in Carteret, N.J., in the 1980s, the yearly flow through of copper scrap and new “blister” copper from Amax’ mines reached 250,000 t per year. The smelter produced from anode slimes 100,000 troy oz of gold, 10,000 troy oz of platinum and 1,000 troy oz of rhodium a month! To the best of my knowledge, no selenium or tellurium was recovered commercially. It was there; it just wasn’t deemed to be of sufficient economic value to repay the costs of separating and purifying it.

Today, the highest grade ores from which copper is produced are being exhausted, and new processes for extracting copper from lower grade ores are now used more and more in places like Bingham Canyon, Utah, where Kenencott Copper, a unit of Anglo-Australian Rio Tinto, operates America’s largest copper mine. A Kenencott operations manager told me that solvent leach operations at Bingham Canyon actually recover more molybdenum from the lower grade ores than electro-winning ever did from the higher grade ores. As for tellurium, the solvent leach technology does not lend itself to the recovery of selenium or tellurium, so unless the prices for selenium and tellurium were to rise dramatically, they would essentially not be recovered.  Today they are thus produced in much less quantity than when the mine, and any other mine of the same type, solely utilized electro-winning of higher grade ores.

As I previously mentioned, global tellurium production was estimated in 1997 by the NREL, to be a maximum of 800 t out of a total possible recovery of 1600 t. Unfortunately, the NREL credibility has been compromised by a more recent NREL publication, “Will we have enough materials for energy-significant PV Production?”. It contains the following paragraph, the statements and conclusions of which are wrong because they are vastly oversimplified:

“In brief, our conclusion is this: Producing 20 GW/year of PV in the United States by 2050 would not create problems with materials availability. Issues surrounding the availability of PV materials at this level simply do not exist. Only indium and tellurium remotely approach becoming bottlenecks at this annual production rate, and simple strategies exist that would solve these problems, including extracting them from ores that are currently mined but unused”.

Many papers have been written in the last few years assuming that all of the tellurium that could be recovered will be recovered, and that copper production will simply continue to increase, thus increasing the production of tellurium.

Both assumptions are questionable. The recovery of copper only from high grade ores is ending and recovery from lower grade ores is more expensive. Also it must be noted that 2008 was the highest production year in history for all metals. From the beginning of the age of the electrification of the West (and of the use of brass-cased ammunition for repeating firearms), until now, has been a period of more than 100 years; the key metal in that electrification, and the manufacturing of brass-cased ammunition, has been copper, and in 100 years, humanity has reached the production level of 16 million t per year.

Let’s assume that the doubling of copper production is possible. It will have to be, ultimately, from lower grade ores, which are abundant, and that, plus the ever-rising cost of energy, will sharply reduce that part of the total that is produced by electro-winning, so it will most likely also sharply reduce the production of tellurium. For argument’s sake, let’s assume that the global production of copper can and will double within 25 years. Note that from 2007 to 2008, the total global production of copper rose just 2%, illustrating that when production levels get to this volume, increases will ordinarily only be marginal.

Let’s also assume that the USGS estimate of global tellurium production in 2007, 500 t, in fact represented only 50% of total possible production, and that tellurium production will keep pace with copper production increase and that all of it will be recovered from the copper in 2034. This means that a total production of 2000 t of tellurium is possible in 2034.

Let’s further assume that all of the tellurium mined in 2034 goes to First Solar to make cadmium telluride thin-film photovoltaic cells.

An article in Wikipedia about cadmium telluride PV cells states that “One gigawatt (GW) of CdTe PV modules would require about 66 tonnes [of tellurium] (at current efficiencies and thicknesses).” The installed capacity for electricity generation in the United States is today, in 2009, about 1,100 GW.

According to the Energy Information Administration publication “Electric Power Annual” for 2007, total installed photovoltaic generating capacity in the United States on 2007 was 503 MW at 38 installations for solar energy conversion – not all of which were based on photovoltaics. This means that in 2007, the US capacity for electrical energy generation that could be produced by solar energy conversion was a maximum of 0.05%.

The USGS report on tellurium, in its discussion of end-uses, states that:

“First Solar Inc. (Phoenix, Ariz.) was the leader in CdTe production, with plants in Ohio and Germany and another plant opening in Malaysia in 2008. In 2007, with an annual capacity of 210 MW, First Solar accounted for 90% of global CdTe cell capacity. By 2010, it was projected that global CdTe cell production capacity will reach 608 MW (Ullal and von Roedern,)”

A recent article entitled “Sustainable Energy”, makes the following point:

“For example, if solar energy is to expand from its present contribution of less than 1% of renewable power generated here last year to, say, 10% of our total power supply, the use of an ingredient in proportions as small as a hundred grams per kilowatt of capacity would translate into a cumulative requirement for tens of thousands of tons. If the substance in question was the Tellurium used in Cadmium-telluride solar cells, its global output would have to expand by at least 10X within a decade or two. That might not be possible, or at least economically feasible.”

In fact, based on 2007 statistics, if the increase were to come from the most scientifically efficient technology, First Solar’s cadmium telluride thin film photovoltaic, the amount of tellurium required would be approximately 220 times the amount for which First Solar had the need and capacity in 2007.

Assuming that the Wikipedia figure of 66 t of tellurium per GW required by First Solar’s technology is accurate, this means that First Solar needed 13 t of tellurium in 2007. But for its solar energy conversion technology to get to a total of 10% of our American 2008 installed capacity, would then require 7,260 t of tellurium! First Solar is predicted to get to a manufacturing capacity of 608 MW in 2010; this means it will need 39 t of tellurium in 2010.

My conclusion from these facts is that First Solar is most likely struggling to obtain tellurium even now, because its demand just for next year could be as much as one-third of the annual global production of tellurium if we believe the most conservative of figures above. In any case, it will be no less than 8% of the USGS guesstimate of total annual global production of tellurium based on the continued production of 16 million t per annum of copper. It cannot be overemphasized that First Solar’s demand for tellurium is a demand for new material. Its technology has created a new demand and does not replace an older demand for tellurium, so we are speaking of additional recovery of tellurium to satisfy First Solar’s requirements.

It may be possible to increase the percentage recovery of tellurium from copper, but it is more likely that the annual global production of tellurium will decrease as copper production technology, even if the total copper produced is growing slightly, moves to solvent leach extraction processes to beneficiate lower grade ores.

First Solar’s problem is not simply one of getting tellurium, but in getting the amount it needs annually. First Solar’s end use product has a long life and can probably be recycled, so that in theory given enough time First Solar could obtain 7,250 t of tellurium and with it manufacture cumulatively enough solar cells to generate 110 GW of electricity, 10% of our current installed capacity from all types of generating devices and fuels.

No one knows even within 50% how much tellurium is produced annually today, but I am going to make the conjecture that it is between 500 and 1,000 t. It is possible that this range could be increased, by increasing the amount of copper processed by electro-winning for tellurium, but it is unlikely that such an agenda would be able to be carried out. Just to demonstrate the economics, it is only necessary to consider that in 2008, copper reached a price of almost $9,000 per t, so that the copper market for 2008 had a high value of $144 billion. 1,000 t of tellurium at its high point in 2008 was worth $200 million. The copper mining industry cannot logically continue to use electro-winning and thus decrease production and raise the costs of producing copper, simply to add even double the $200 million that doubling the tellurium output would bring, in the face of the reality that such an undertaking could cost billions of dollars of copper production. There is no economic driver for increasing the production of tellurium!

I will give more detail on this topic in future articles and my forthcoming book, but suffice to say that for First Solar to increase its production by 2 GW per year, which is three times its projected manufacturing capacity for 2010, which in turn will be three times its 2007 capacity, will require a supply of 120 t a year of tellurium, and that at that enormous production rate, it would take 60 years to achieve the changeover of 10% of the installed electric generating capacity of the USA in 2008, to the cadmium telluride thin flim photovoltaic cells today produced by First Solar.

Solar cells today are not economical; they cost more per watt of output than fossil fuel production of electricity. Solar cells today are only manufactured and sold through taxpayer subsidies. The Cap and Trade promoters have overlooked the increase in costs their regime will bring to the metal mining and refining industry. Whatever economies of scale have been “conjectured” by First Solar, will now be obsolete and the time when their product could be competitive with fossil fuel, production will move further into the future.

The biggest risks of all is that non-solar industries will raise their demand for tellurium and both domestic and foreign other solar cell makers will increase their demand for tellurium. The United States produces today about 10% of the world’s copper, which means that it likely produces 10% of the world’s tellurium. It is likely that First Solar’s current production demand, in the United States, alone requires that the US industry become reliant on foreign sources of tellurium already. In that case, because US production of copper can increase only marginally, if First Solar is to grow it will certainly have to do so by creating jobs in mining overseas or by moving its production overseas. Other sustainable energy companies chasing subsidies and “stimulus” funds, will surely complain that subsidizing First Solar, is simply subsidizing foreign resource producers.

I am not qualified to analyze First Solar’s balance sheet to see to what level their current costs would have to do, to make their products competitive with other electric generating technologies of all types. However, I am qualified to predict that the cost of tellurium will increase and its availability will decrease in the near future.

A company such as First Solar, which is critically dependent on a secure supply of tellurium to exist and on an unsustainable growth in the supply to it of tellurium for it to grow and achieve competitive pricing is a big risk for short-term investors. The maximum supply and production levels attainable of tellurium are quantifiable even if the actual production figures are murky, and they do not bode well for the future of First Solar if it must make profits to survive.