Sunday, November 29, 2009

Small-Scale Manufacturing and Digital Fabbers - The Question of Electronics

One of the consequences of the decline of available fossil-fuel energy is the contraction of our large-scale, global industrial economy. The decline in supplies of fossil fuels will make globalism prohibitively expensive as time goes on, due to the ever-increasing energy cost of shipping bulk-manufactured goods thousands of miles from their point of manufacture to their point of final sale. Many elements of modern society will therefore only survive via the revival of local, small-scale manufacture of goods.

The creation of small-scale, do-it-yourself digital fabricators (referred to from here on as “fabbers”), has been promoted as a key to the revival of modern-day small-scale manufacturing. According to many fabber proponents and enthusiasts, the rise of fabbers promises to do for manufacturing what inexpensive consumer entertainment electronics did for the creation of media. Whereas cheap consumer electronics enabled everyone to be a potential creator of art, education or entertainment, fabbers might enable everyone to be a potential creator of useful manufactured goods.

But for fabbers to serve as a true long-term solution to the breakdown of centralized industrial production, they must be able to create everything needed for sustainable localized economies – including parts to make more fabbers. To the extent that the making of fabbers requires parts or components that can only be made by large-scale plants in today's economy, to that extent fabbers are not really sustainable. One item of concern is thus the microelectronic components used to control fabbers, as these microelectronic controllers are now made in large, energy-intensive semiconductor chip plants. There are many issues of concern for those who want to try making microprocessors on a small scale, such as the very demanding and exacting conditions required for manufacture (vacuum chambers, ultrapure materials and clean rooms), and the energy required to achieve these conditions.

These conditions apply to all semiconductor-based microelectronics, though their impact varies depending on whether we are considering organic or inorganic semiconductor materials. Today's post will consider manufacture of inorganic semiconductor microelectronics. In this post, I do not promise to come to definite conclusions, but rather to raise important questions. It seems to me that these questions are too often not addressed by those who enthusiastically promote a “fabber revolution” as a solution to economic collapse. My posts on this topic are designed to provoke a conversation on this subject. There are four questions which I'd like to see addressed:

The Question of Energy

Almost all semiconductors in use at present are inorganic. (Liquid-crystal displays, some flat-panel screens and some RFID tags are notable exceptions.) Most inorganic semiconductor electronics are silicon-based.

In its natural form, silicon is literally dirt-cheap. However, the silicon found in sand and dirt is not nearly pure enough for use in high-speed electronics. The process of purification is not nearly as cheap. Metallurgical grade silicon (98 percent pure and above) is created by the reaction of high purity silica with other materials in an electric arc furnace heated to over 1900 degrees C. A method also exists for extracting pure silicon (purity greater than 99 percent) from silica by molten salt electrolysis. But this process also requires high temperatures (around 900 degrees C).

Electronic-grade silicon must be millions of times purer than 99 percent pure. The processes of this purification start with the aforementioned metallurgical grade silicon as a feedstock. They are all very energy-intensive, with the Siemens process (Chemical Vapor Deposition) having the highest energy requirement. Getting from beach sand to electronic-grade silicon is not cheap!

Once the silicon is at the right purity, it must be doped with trace elements in order to produce the desired semiconducting properties. This process is also energy and equipment-intensive, and requires a vacuum chamber containing pure silicon rods heated to 1000 degrees C. Many of the dopant chemicals are extremely poisonous, and some are also explosive.

Once the properly doped silicon has been created, it is cut into wafers which are etched and deposited with other dopants and contact metals in vacuum chambers in order to make the final microelectronic chips used in almost all modern digital devices. The processes of this manufacture are all quite expensive, both in labor, capital (machinery required) and energy. Modern digital devices are as cheap as they are simply because not much semiconductor material is needed anymore in order to make chips of great computational power. Yet energy is generally becoming more expensive as time passes, and shortages of dopant materials are also beginning to appear.

The Question of Dopants (And Other Exotic Materials)

The dopants used to alter the conducting properties of silicon and other semiconductors are themselves hard to find, hard to mine and relatively scarce in many cases. Antimony is one such dopant, used for both silicon and germanium semiconductors, and it has found extensive use in newly developed rewritable memory for digital devices. Most of the remaining antimony in the world is produced by China, and there is no U.S. domestic antimony production. Gallium is another material whose manufacturing users experienced a recent shortage, as was the case with indium also. Thallium is yet another metal whose supply has become constrained at nearly the same time that demand for the metal has increased. Many dopants and other industrial metals have witnessed Hubbert production peaks and are now in decline.

It may be that the electronics industry will experience a dead end in the use of certain elements within the next few decades, as the available supplies of these elements run out. This will mean a stop to the making of microelectronics that depend on these elements for doping. If continued advances in electronics are to continue, the industry will have to find alternatives to expensively produced inorganic semiconductors doped with scarce materials.

Hope On The Horizon? (The Promise of Exotic Materials)

Within the last few years, there have been exciting announcements of the discovery of exotic forms of common materials, forms whose properties hold the possibility of creating wonder microelectronics which don't need exotic dopants. One such development is the creation of silicon nanotubes, which have recently been fabricated into dopant-free nanotransistors by crossing the nanowires over each other and adding tiny metal contacts known as “Schottky contacts.”

However, the creation of these exotic nanowires requires a correspondingly exotic process. The first step is the production of silane from metallurgical grade silicon at a temperature exceeding 300 degrees C. The resulting silane is pyrophoric and explosive, and must be carefully handled. Then the silane is passed over a metal catalyst in a special chamber heated to at least 400 degrees C. This step is what produces the silicon nanowires. While the process can yield nano-transistors and other nano-components that do not require dopants, the process itself is still quite energy-intensive. One publication states that the silicon nanowire breakthrough may lead to “printable electronics” that can be produced by an inkjet printer. I myself am a bit skeptical. If someone could kindly explain to me how this would work, I would happily listen.

Concluding Questions:

The promoters of one particular fabber project state that their concept is the key to “wealth without money,” and that a society supplied by fabbers can “create wealth with a minimal need for industrial manufacturing.” They even talk of a society that is able to provide its own stocks of raw materials by turning crops into polymer feedstocks for fabrication by their fabbers, so that a cycle of wealth could be perpetuated (while reducing greenhouse gas emissions at the same time – a neat bargain!).

I remain unconvinced (but not dogmatically so). I think that, at least as far as energy and the resource-intensive microelectronics needed to run these fabbers, their promoters have overlooked the effects of looming scarcity, and the difficulties posed by the breakdown of our present industrial society. Has anyone made a do-it-yourself garage fabber that can make silicon nanowires? How about a DIY garage fabber that can even make metallic silicon? Are there fabbers that can make high-quality vacuum deposition chambers? Are there fabbers that can dope pure silicon without the risk of toxic gases leaking out and poisoning a few households in a neighborhood? Has anyone rigorously addressed the problem of obtaining large supplies of metallic silicon in an energy-constrained future? (This is the BIG question.) Most importantly, how much energy will all of this take? How long will we retain access to that kind of energy? The future I envision for electronics looks rather different from that of the optimists, but I would welcome further discussion and enlightenment on this subject, including some more rigorous numerical analyses.

The next time I address this topic, we will consider organic (polymeric) electronics. Stay tuned...


No comments: