DETROIT () -- It's difficult for me to imagine two more completely opposed situations of fundamental demand and supply than those for the chemical elements, silicon and tellurium. Silicon is the second most abundant element in the earth's crust, according to the U.S. Geological Survey; its finely divided oxide derived from the weathering of the earth's ubiquitous silicate rocks and quartz over billions of years makes up the overwhelming majority of the sand on the world's beaches and in its deserts. Silicon dioxide, sand, is accessible and present on the earth's surface in a literally inexhaustible supply.
But sand is not the source of silicon metal, the precursor of the ultra high purity 'electronic' grade silicon, which is required for the creation of silicon photovoltaic solar cells, which generate electricity when ordinary sunlight falls on them, and which when wired together in sufficient numbers can be used to power devices far from the nearest coal, oil, or nuclear powered electricity supply grid.
Tellurium, which I will discuss next time, is rare, even in the earth's crust, and is only today recovered, almost entirely, as a byproduct of copper mining. There is not, in any case, enough of it produced annually to support a major photovoltaic solar cell industry based on a transition metal, such as cadmium, telluride film technology with the goal of making a significant contribution to global energy demand.
Silicon metal, from which the ultra-pure silicon used for photovoltaic cell construction is ultimately derived, is made in a process that begins with the minerals, quartz or quartzite, which are relatively pure forms of crystalline silicon dioxide formed either under extremes of temperature or pressure or both over, literally, eons of time. In order to reverse this process, i.e., to 'chemically reduce' the silicon dioxide to silicon and to melt, it takes the intense application of heat in the presence of a 'chemical reducing agent,' in a situation where the chemical reaction does not reach a balance until all of the silicon dioxide has been 'reduced' to silicon metal.
In highly industrialized countries this reaction is accomplished by mixing, say, quartz with a pure form of carbon, such as from cleaned coal or graphite, and striking an electric arc between a pure graphite anode beneath the surface of the quartz/carbon mixture and a conducting crucible which contains the reactants. The intense heat of the electric arc soon melts some of the quartz and as the temperature of the melt rise the carbon attracts the oxygen from the silicon and forms the gas, carbon dioxide, which is continuously allowed to leave the reactor thus driving the reaction to continue until all of the carbon or all of the quartz has been consumed.
As the USGS data linked above shows, there is no practical way to measure the supply of quartz and quartzite in the U.S. or anywhere else; the supply is, for all intents and purposes, inexhaustible.
About 99.9% of the silicon metal produced above is used directly or processed further to make ferrosilicon, a pre-cooked form of silicon that is directly soluble in molten iron and some other important industrial metals, in small quantities, and is the most efficient (i.e., cost effective) way to add silicon for alloying.
By now you must be wondering what all the fuss about a shortage of silicon to make photovoltaic cells for solar power generation can possibly be about.
It is about money, i.e., the enormous cost of producing electronic grade (99.9999+) silicon from the mass produced pure silicon metal made in a submerged arc electric furnace. A typical purifying process takes a relatively small quantity of the silicon metal made in the electric furnace, which has been selected as relatively pure, sometimes as good as 99+%, and reacting it with chlorine gas (derived commercially in vast quantities during the production of industrial mainstay chemical sodium hydroxide, known as caustic soda. The major use for chlorine is water purification.
The silicon tetrachloride produced by the above mentioned reaction is then itself purified by distillation thus separating the silicon tetrachloride from other volatile chlorides produced when the chlorine reacts with the impurities in the silicon. When repeated distillations have produced as pure a silicon tetrachloride as possible the silicon tetrachloride is then chemically reduced (perhaps by heating it with very pure hydrogen, for example) to high purity silicon metal, perhaps 99.99+% pure. This relatively impure, from the point of view of electronic uses, silicon is now placed in an ultra-pure non reactive high temperature stable 'boat' or simply, in the form of a crude rod, clamped at both ends and a thin hot disc shaped zone is created by a circumferential induction coil at one end of the rod, hot enough to dissolve impurities (remember hot water dissolves more salt than cold water) and large enough to hold, dissolved, all the impurities in the total rod.
When the very slow moving 'zone' reaches the other end of the rod, the heater ring turns off and returns on its track to the beginning of the rod, where the process starts again. After a number of passes one end of the silicon rod is much purer than the other. At some predetermined point the 'dirty' end is cut off and the pure end recovered and handled very carefully as it has become single crystal ultra pure electronic grade silicon with a selling price of thousands of dollars a kilogram; it must not be touched by bare hands with their sodium laden sweat, nor exposed to air with its reactive acid forming oxides, such as those of carbon or nitrogen, much less the particulate contaminants found in ordinary air.
The ultra-pure single crystal silicon rod now goes to a chip maker who slices it first into wafers and then, after a complex process, involving imaging, etching, the deposition and removal of many layers, and doping, i.e. recontamination of the silicon in selected places, slices the wafers into chips upon which the integrated circuits have been constructed, by the steps mentioned, and more, for electronic uses.
For most photovoltaic solar cell applications a slightly lower grade of silicon wafer can be used, a polysilicon one, which means that it has been cut from a rod that was not processed so as to become just one single crystal, but rather remains polycrystalline. This polysilicon is less expensive than single crystal silicon, but not all that much, since it is still very, very, pure.
Silicon PV cells can also be made as thin films but the starting material for that process must usually be ultra pure polysilicon which is then, in one process, made back into a volatile silicon compound, such as silicon tetrachloride, and reacted on the (conducting) substrate upon which it is to be deposited with a reducing agent to produce a polysilicon film on that surface. These films are then carefully re-contaminated; the process is called 'doping', with carefully selected metal ions to enhance their solar conversion efficiency. One such process actually is done so as to produce non-crystalline (amorphous) silicon films, which are then doped with hydrogen.
The point of all of this explanation is to show you that silicon metal is relatively cheap and the supply of raw material from which it is made is basically inexhaustible. It is the polysilicon, from which PV silicon solar cells are made, the supply of which is limited to the number of ultra purifying installations in existence and the time it takes to completely process commercially pure silicon metal produced in submerged arc electric furnaces all the way to ultra pure polysilicon that is the materials bottle neck for expanding the use of silicon for PV cells.
Ultra purification techniques for electronics' manufacturing materials, such as silicon, were pioneered in Europe and developed in the U.S. Building an installation and bringing it on line to budget and on time is only possible, the first time, if you have done it before. This means that no matter how many books you have read, exams you have passed, or whether or not you have the other guys blueprints and notes you are not going to undertake the construction of such an installation lightly.
The silicon PV solar cell industry is behind the same eight ball as the hybrid car battery industry; which is now mass producing and is ready to increase production of an existing technology, nickel metal hydride batteries, just when the end-user industry has decided it wants only an unproven but newer, sexier, technology, lithium-ion batteries. Just as the silicon based PV solar cell industry got going commercially it was announced that silicon was pass'e and that cadmium telluride based thin film devices were much more efficient-true- and much easier and cheaper to make-false. Thus, since the silicon ultra pure rod makers are in no hurry to increase production in a world facing a consumer product melt-down in the purchase of electronics, and, additionally, they seem resistant to downgrading their facilities to make polysilicon, there is a road block; those who know how to make polysilicon are reluctant to make the investment especially as it will only benefit a small industry and tie up resources that would be needed if electronics, a huge industry, rebounds.
Do I need to mention the capital or the non-existent credit facilities that would have to be drawn down to build polysilicon capacity?
I think that it will turn out that PV solar cells are not the answer to the world's needs for less polluting sources of electric energy, but only a small part of the solution. One problem is that it takes a lot of chemical processing to produce a kilogram of silicon from which to ultimately make a PV solar cell, and this uses and produces an enormous amount of carbon dioxide, carbon monoxide, chlorine and so forth. Someone needs to do the calculation to see just how much electricity a silicon PV solar cell would have to produce, without producing pollution, to overcome the energy needed to manufacture it and to simultaneously overcome the pollution caused by the total of the manufacturing processes required to produce it.
In the interests of full disclosure and of telling a good story I need you to know that I began working on photovoltaic cells in the mid 1960s when I was employed to work out methods to prepare (then) exotic materials and measure their electronic properties, including any photo-ohmic (change of resistance when exposed to light) or photovoltaic ones.
I was preparing thin films of many tellurium containing materials, and the PV properties of some of them were standing out, but the information was not followed up upon at the time, because Stanford Ovshinsky, my boss, had come across the fact that when high purity amorphous silicon films were doped with hydrogen, they became strongly photovoltaic. He actually started the mass production of thin film PV amorphous silicon solar cells, shortly after the discovery, and I think it was the very first attempt to mass produce thin film large size cell groupings commercially. I was involved with the ultra purification of materials at the time and so I learned the processes by which materials such as silicon and germanium were prepared and purified. It was the at the very beginning of the commercial development of the integrated circuit, and I observed the work at Fairchild, Motorola, Intel, ITT and even, later on, Elkem of Norway's Silicon Metal Plant in West Virginia and their operations in Niagara Falls, New York making graphite electrodes for arc furnaces and using them to make ferroalloys.
I well remember sitting across from (later, Sir) Neville Mott, in the late 1960s, as he discussed the ideas that led him to the explanation of the photovoltaic properties of amorphous silicon-hydrogen doped solar cells. Mott was awarded the Nobel Prize in physics for 1975.
I took this short trip down memory lane to let you know that I have more than a little background in this topic, and I want to say that investing in silicon based PV solar cell technology is too risky for the small investor. The economics do not work on a large scale.
Monday, I will continue with a topic that I think I really know a lot about, tellurium based PV solar cells in an article entitled, "Materials for Solar Photovoltaic Cells II: Supply and Demand Cast a Shadow on Solar, Speaking of Tellurium."