6 Feb
The stuff I want to write about is silicon, and my intention is to go through the process of gathering, synthesizing and composing in a public sort of way, recording the process on this page so that I can know what I think by seeing what I say.

Why silicon? If my 'project' has to do with technologies of memory, then the substance which is at the heart of today's foremost memory technology (the computer) seems particularly apposite; I need to get my own thoughts in order, need to marshall facts and questions, need to sort out for myself what I understand and what I allow my mind to slide over --partly a matter of recovery of things I've read or learned in the past, partly a search for new material to synthesize the bits I have into something more satisfactory, more comprehensive.

For a good 8 years or so I've spoken jokingly of computers as 'silicon-based life forms', a conceit which goes back to something I read (I know not where, and wish I could recover the original source --but see this for one other user of the term) that said that carbon-based life forms were one possibility, but that silicon was another logical possibility, sort of an evolutionary path not taken by terrestrial life. One could argue that the greatest accomplishment of the carbon-based life form we are is (?will be?) the creation of a silicon-based life form. This verges upon the religious, and emerges directly from the book of the moment, picked up yesterday from the New Books shelf in the Science Library: Ray Kurzweil's The Age of Spiritual Machines [Q335 .K88 1999]. This book can be read as sci-fi, as technology-based extrapolation, as philosophical statement, as religious tract... and it's not clear to me just yet how I'm reading it. But consider these passages:

Evolution has been seen as a billion-year drama that led inexorably to its grandest creation: human intelligence. The emergence in the early twenty-first century of a new form of intelligence on Earth that can compete with, and ultimately significantly exceed, human intelligence will be a development of greater import than any of the events that have shaped human history. It will be no less important than the creation of the intelligence that created it, and will have profound implications for all aspects of human endeavor, including the nature of work, human learning, government, warfare, the arts, and our concept of ourselves. (page 5)

...the next stage of evolution should measure its salient events in mere thousands of years, too quick for DNA-based evolution. This next stage of evolution was necessarily created by human intelligence itself, another example of the exponential engine of evolution using its innovations from one period (human beings) to create the next (intelligent machines). (page 35)

Silicon is at the very base of all human technology, by virtue of the fact that stone (specifically igneous rock, but also therefore metamorphic and sedimentary rock) is "basically an oxide of silicon" (Flanagan's Version, pg 192). On the cosmic level silicon also plays a part:

When the first stars lit up, they consisted of hydrogen and helium. Their nuclear energy came from the fusion of hydrogen nuclei into helium nuclei. What happens when a star's supply of hydrogen runs out? It swells up into the large cool star called a red giant. At the same time the core of the star, the place where it is hot enough and dense enough for nuclear reactions to proceed, shrinks and gets still hotter and denser. Under those conditions the energy of the helium nuclei is sufficient for them to fuse.

In the process helium nuclei are built up into the nuclei of carbon and oxygen. What happens when the red giant's supply of helium runs out? It first begins to burn the carbon and then begins to burn the oxygen. In the process the carbon and oxygen nuclei are built up into still heavier elements. The switchover from one nuclear fuel to another as each runs out continues up the periodic table of the elements through silicon. When it gets to elements in the vicinity of iron, it stops. The fusion reactions that could make elements heavier than iron do not release energy, they absorb it. (Flanagan's Version pp. 72-73)

Here's some basic information, culled from Macmillan Encyclopedia of Chemistry:

...the second most abundant element in the Earth's crust, following oxygen. The name is derived from the Latin word silex, meaning "flint"... Silicon's chemistry is that of the covalent bond, with the major, naturally occurring compounds being either oxides or silicates...sand, quartz, amethysts, agates and opals... clay, mica, granite, asbestos... (pg. 1345)

Kirk-Othmer gives some other bits that need to be integrated (which, in this instance, means I have to understand what they mean...):

Electrical properties. Silicon is a semiconductor with a band gap Eg of 1.12 eV at ca 25 C; Eg is the amount of energy required to raise an electron from the valence band to the conduction band.
(this band gap allows operation up to about 200 C, whereas germanium is limited to about 90 C [Shimura 1989:1]

Chemical properties. Silicon, carbon, germanium, tin and lead comprise the Group IVA elements of the periodic system. Silicon and carbon form silicon carbide which, although most widely known as an abrasive and heating element, is also a semiconductor. Germanium and silicon are isomorphous and thus are mutually soluble in all proportions. Molten tin and lead are immiscible with molten silicon. Molten silicon dissolves most materials and no container material has been found that is not noticeably dissolved.

The elements in the adjacent columns (IIIA and VA) form compounds with silicon and also enter substantially in small amounts into the lattice of a silicon crystal.

Oxygen forms strong bonds with silicon. There are two oxides, numerous silicates, and almost endless varieties of silicones... (p. 1057)

Reduction. Semiconductor-grade silicon is prepared by careful purification of some easily reduced silicon compounds such as SiCl4 followed by reduction with an equally pure reducing agent (usually hydrogen). (p. 1058)

The above doesn't really answer a lot of questions (how do they actually make semiconductor-grade silicon? how do silicon chips actually work? etc.), but answers to those questions are not all that hard to find:
The importance of silicon in the electronic device technology relies on semiconducting properties, which can be controlled by small additions of impurity elements, that is, dopants, such as Group III elements (e.g. boron) for acceptors and Group V elements (e.g. phosphorus) for donors... [Shimura 1989:58]

Here's a link to a chemical fact sheet for silicon.

A page on the crystal structure of silicon helps a bit, and introduces n- and p-type materials, essential to an understanding of transistors. This is part of a site at University of New Mexico which has a summary of Integrated Circuit fabrication in linked pages, including one on wafer manufacture. A lot of the essentials I've been hunting for are well presented in these pages.

There'a VLSI tutorial from a Swiss source which covers some of the same territory, and details the Czochralski Process for growing silicon crystals, and a commercial site describes the process nicely.

The evolution of chips is a diverting part of this subject, and is covered in a number of places. Elegant pictures (to 1983) can be found in

 AUTHOR       Augarten, Stan.
 TITLE        State of the art : the brilliant career of the integrated circuit
 PUBLISHER    New Haven : Ticknor & Fields, 1983.
 SUBJECT      Integrated circuits -- Popular works.
 Science Library        TK7874 .A84 1983
This book includes pictures of (a re-creation of) the first transistor (1947) and integrated circuit (1958) --both of which used germanium, not silicon-- and ends with the still-experimental gallium arsenide and Josephson junction chips that were the frontier in 1983.

I did a web search to update that, and found several things that look worthwhile: Evolution of the microchip (Menelaos Levas), and (at the same UK site) Making use of Moore's law: How computer architects have used more transistors to improve performance (Stavros Polyviou)

(Moore's Law turns up in all sorts of places. Gordon Moore, one of the founders of Intel, made the generalization in 1973 that the number of transistors on a chip doubles every 18 months. He's revised the prediction to (I think) 21 months, but it's held up remarkably well as chips have moved from thousands to millions of transistors --essentially a matter of progressive miniaturization technology in chip fabrication)

And (in the Levas article bibliography) I ran across a recent article that turns out to be available in full text: Silicon trends and limits for advanced microprocessors (Mark Bohr). This one gives a good basis for comprehending the current state of the art of miniaturization, the next few steps on the horizon, and the reasons for ultimate limits of miniaturization of integrated circuits on silicon substrates. Here's the essence:

Patterns are printed on [silicon] integrated circuit wafers by focusing light through masks and onto photosensitive films on the wafer surface. This technique, called optical lithography, has constantly been improved... By reducing the wavelength of the exposing light, using better lenses, and improving the photosensitive films, optical lithography is now patterning 0.25 micrometers minimum feature sizes in high-volume production. (Bohr 1998:81)

Another book with breathtaking images is

 AUTHOR       McCarty, Cara.
 TITLE        Information art : diagramming microchips / Cara McCarty.
 PUBLISHER    New York : Museum of Modern Art : distributed by Harry N. 
Abrams, c1990.
 NOTE         "This exhibition is made possible by the Intel Corporation
                Foundation"--T.p. verso.
 SUBJECT      Integrated circuits -- Design and construction.
              Integrated circuits -- Exhibitions.
 Leyburn Library        FOLIO NX458 .M35 1990

And here's a brand new book that deals with environmental and safety issues in the semiconductor industry:

 AUTHOR       Mazurek, Jan, 1965-
 TITLE        Making microchips : policy, globalization, and economic
                restructuring in the semiconductor industry / Jan Mazurek.
 PUBLISHER    Cambridge, MA ; London : MIT Press, c1999.
 SUBJECT      Semiconductor industry -- Employees -- Health and hygiene.
              Integrated circuits industry -- Employees -- Health and hygiene.
              Computer industry -- Employees -- Health and hygiene.
 Leyburn Library        HD7269.S44 M39 1999

Browsing in this one led me to the Silicon Valley Toxics Coalition web page, and thus to the general questions of costs and benefits and the old cui bono? [literally, 'to whom the good?'] that needs to be factored into any historical consideration of technology.