by Sam Kean
In fact, scientists nowadays can decode vastly longer molecules than “acetyl… serine.” The current record is a gargantuan protein whose name, if spelled out, runs 189,819 letters. But during the 1960s, when a number of quick amino acid sequencing tools became available, scientists realized that they would soon end up with chemical names as long as this book (the spell-checking of which would have been a real bitch). So they dropped the unwieldy Germanic system and reverted to shorter, less bombastic titles, even for official purposes. The 189,819-letter molecule, for instance, is now mercifully known as titin.* Overall, it seems doubtful that anyone will ever top the mosaic virus protein’s full name in print, or even try.
That doesn’t mean aspiring lexicographers shouldn’t still brush up on biochemistry. Medicine has always been a fertile source of ridiculously long words, and the longest nontechnical word in the Oxford English Dictionary just happens to be based on the nearest chemical cousin of carbon, an element often cited as an alternative to carbon-based life in other galaxies—element fourteen, silicon.
In genealogy, parents at the top of a family tree produce children who resemble them, and in just the same way, carbon has more in common with the element below it, silicon, than with its two horizontal neighbors, boron and nitrogen. We already know the reason. Carbon is element six and silicon element fourteen, and that gap of eight (another octet) is not coincidental. For silicon, two electrons fill the first energy level, and eight fill the second. That leaves four more electrons—and leaves silicon in the same predicament as carbon. But being in that situation gives silicon some of carbon’s flexibility, too. And because carbon’s flexibility is directly linked to its capacity to form life, silicon’s ability to mimic carbon has made it the dream of generations of science fiction fans interested in alternative—that is, alien—modes of life, life that follows different rules than earth-bound life. At the same time, genealogy isn’t destiny, since children are never exactly like their parents. So while carbon and silicon are indeed closely related, they’re distinct elements that form distinct compounds. And unfortunately for science fiction fans, silicon just can’t do the wondrous tricks carbon can.
Oddly enough, we can learn about silicon’s limitations by parsing another record-setting word, a word that stretches a ridiculous length for the same reason the 1,185-letter carbon-based protein above did. Honestly, that protein has sort of a formulaic name—interesting mostly for its novelty, the same way that calculating pi to trillions of digits is. In contrast, the longest nontechnical word in the Oxford English Dictionary is the forty-five-letter “pneumonoultramicroscopicsilicovolcanoconiosis,” a disease that has “silico” at its core. Logologists (word nuts) slangily refer to pneumonoultramicroscopicsilicovolcanoconiosis as “p45,” but there’s some medical question about whether p45 is a real disease, since it’s just a variant of an incurable lung condition called pneumonoconiosis. P16 resembles pneumonia and is one of the diseases that inhaling asbestos causes. Inhaling silicon dioxide, the major component of sand and glass, can cause pneumonoconiosis, too. Construction workers who sandblast all day and insulation plant assembly-line workers who inhale glass dust often come down with silicon-based p16. But because silicon dioxide (SiO2) is the most common mineral in the earth’s crust, one other group is susceptible: people who live in the vicinity of active volcanoes. The most powerful volcanoes pulverize silica into fine bits and spew megatons of it into the air. Those bits are prone to wriggling into lung sacs. Because our lungs regularly deal with carbon dioxide, they see nothing wrong with absorbing its cousin, SiO2, which can be fatal. Many dinosaurs might have died this way when a metropolis-sized asteroid or comet struck the earth 65 million years ago.
With all that in mind, parsing the prefixes and suffixes of p45 should now be a lot easier. The lung disease caused by inhaling fine volcanic silica as people huff and puff to flee the scene is naturally called pneumono-ultra-microscopic-silico-volcano-coniosis. Before you start dropping it in conversation, though, know that many word purists detest it. Someone coined p45 to win a puzzle contest in 1935, and some people still sneer that it’s a “trophy word.” Even the august editors of the Oxford English Dictionary malign p45 by defining it as “a fractious word,” one that is only “alleged to mean” what it does. This loathing arises because p45 just expanded on a “real” word. P45 was tinkered with, like artificial life, instead of rising organically from everyday language.
By digging further into silicon, we can explore whether claims of silicon-based life are tenable. Though as overdone a trope in science fiction as ray guns, silicon life is an important idea because it expands on our carbon-centric notion of life’s potential. Silicon enthusiasts can even point to a few animals on earth that employ silicon in their bodies, such as sea urchins with their silicon spines and radiolarian protozoa (one-celled creatures) that forge silicon into exoskeletal armor. Advances in computing and artificial intelligence also suggest that silicon could form “brains” as complicated as any carbon-based one. In theory, there’s no reason you couldn’t replace every neuron in your brain with a silicon transistor.
But p45 provides lessons in practical chemistry that dash hopes for silicon life. Obviously silicon life forms would need to shuttle silicon into and out of their bodies to repair tissues or whatever, just as earth-based creatures shuttle carbon around. On earth, creatures at the base of the food chain (in many ways, the most important forms of life) can do that via gaseous carbon dioxide. Silicon almost always bonds to oxygen in nature, too, usually as SiO2. But unlike carbon dioxide, silicon dioxide (even as fine volcanic dust) is a solid, not a gas, at any temperature remotely friendly to life. (It doesn’t become a gas until 4,000°F!) On the level of cellular respiration, breathing solids just doesn’t work, because solids stick together. They don’t flow, and it’s hard to get at individual molecules, which cells need to do. Even rudimentary silicon life, the equivalent of pond scum, would have trouble breathing, and larger life forms with multiple layers of cells would be even worse off. Without ways to exchange gases with the environment, plant-like silicon life would starve and animal-like silicon life would suffocate on waste, just like our carbon-based lungs are smothered by p45.
Couldn’t those silicon microbes expel or suck up silica in other ways, though? Possibly, but silica doesn’t dissolve in water, the most abundant liquid in the universe by far. So those creatures would have to forsake the evolutionary advantages of blood or any liquid to circulate nutrients and waste. Silicon-based creatures would have to rely on solids, which don’t mix easily, so it’s impossible to imagine silicon life forms doing much of anything.
Furthermore, because silicon packs on more electrons than carbon, it’s bulkier, like carbon with fifty extra pounds. Sometimes that’s not a big deal. Silicon might substitute adequately for carbon in the Martian equivalent of fats or proteins. But carbon also contorts itself into ringed molecules we call sugars. Rings are states of high tension—which means they store lots of energy—and silicon just isn’t supple enough to bend into the right position to form rings. In a related problem, silicon atoms cannot squeeze their electrons into tight spaces for double bonds, which appear in virtually every complicated biochemical. (When two atoms share two electrons, that’s a single bond. Sharing four electrons is a double bond.) Silicon-based life would therefore have hundreds of fewer options for storing chemical energy and making chemical hormones. Altogether, only a radical biochemistry could support silicon life that actually grows, reacts, reproduces, and attacks. (Sea urchins and radiolaria use silica only for structural support, not for breathing or storing energy.) And the fact that carbon-based life evolved on earth despite carbon being vastly less common than silicon is almost a proof in itself.* I’m not foolish enough to predict that silicon biology is impossible, but unless those creatures defecate sand and live on planets with volcanoes constantly expelling ultramicroscopic silica, this element probably isn’t up to the task of making life live.
Luckily for
it, silicon has ensured itself immortality in another way. Like a virus, a quasi-living creature, it wriggled into an evolutionary niche and has survived by preying parasitically on the element below it.
There are further genealogical lessons in carbon and silicon’s column of the periodic table. Under silicon, we find germanium. One element down from germanium, we unexpectedly find tin. One space below that is lead. Moving straight down the periodic table, then, we pass from carbon, the element responsible for life; to silicon and germanium, elements responsible for modern electronics; to tin, a dull gray metal used to can corn; to lead, an element more or less hostile to life. Each step is small, but it’s a good reminder that while an element may resemble the one below it, small mutations accumulate.
Another lesson is that every family has a black sheep, someone the rest of the line more or less has given up on. In column fourteen’s case, it’s germanium, a sorry, no-luck element. We use silicon in computers, in microchips, in cars and calculators. Silicon semiconductors sent men to the moon and drive the Internet. But if things had gone differently sixty years ago, we might all be talking about Germanium Valley in northern California today.
The modern semiconductor industry began in 1945 at Bell Labs in New Jersey, just miles from where Thomas Alva Edison set up his invention factory seventy years before. William Shockley, an electrical engineer and physicist, was trying to build a small silicon amplifier to replace vacuum tubes in mainframe computers. Engineers loathed vacuum tubes because the long, lightbulb-like glass shells were cumbersome, fragile, and prone to overheating. Despise them as they may, they needed these tubes, because nothing else could pull their double duty: the tubes both amplified electronic signals, so faint signals didn’t die, and acted as one-way gates for electricity, so electrons couldn’t flow backward in circuits. (If your sewer pipes flowed both ways, you can imagine the potential problems.) Shockley set out to do to vacuum tubes what Edison had done to candles, and he knew that semiconducting elements were the answer: only they could achieve the balance engineers wanted by letting enough electrons through to run a circuit (the “conductor” part), but not so many that the electrons were impossible to control (the “semi” part). Shockley, though, was more visionary than engineer, and his silicon amplifier never amplified anything. Frustrated after two unfruitful years, he dumped the task onto two underlings, John Bardeen and Walter Brattain.
Bardeen and Brattain, according to one biographer, “loved one another as much as two men can…. It was like Bardeen was the brains of this joint organism and Brattain was the hands.”* This symbiosis was convenient, since Bardeen, for whom the descriptor “egghead” might have been coined, wasn’t so adept with his own hands. The joint organism soon determined that silicon was too brittle and difficult to purify to work as an amp. Plus, they knew that germanium, whose outer electrons sit in a higher energy level than silicon’s and therefore are more loosely bound, conducted electricity more smoothly. Using germanium, Bardeen and Brattain built the world’s first solid-state (as opposed to vacuum) amplifier in December 1947. They called it the transistor.
This should have thrilled Shockley—except he was in Paris that Christmas, making it hard for him to claim he’d contributed to the invention (not to mention that he had used the wrong element). So Shockley set out to steal credit for Bardeen and Brattain’s work. Shockley wasn’t a wicked man, but he was ruthless when convinced he was right, and he was convinced he deserved most of the credit for the transistor. (This ruthless conviction resurfaced later in Shockley’s declining years, after he abandoned solid-state physics for the “science” of eugenics—the breeding of better human beings. He believed in a Brahmin caste of intelligentsia, and he began donating to a “genius sperm bank”* and advocating that poor people and minorities be paid to get sterilized and stop diluting humankind’s collective IQ.)
Hurrying back from Paris, Shockley wedged himself back into the transistor picture, often literally. In Bell Labs publicity photos showing the three men supposedly at work, he’s always standing between Bardeen and Brattain, dissecting the joint organism and putting his hands on the equipment, forcing the other two to peer over his shoulders like mere assistants. Those images became the new reality and the general scientific community gave credit to all three men. Shockley also, like a petty prince in a fiefdom, banished his main intellectual rival, Bardeen, to another, unrelated lab so that he, Shockley, could develop a second and more commercially friendly generation of germanium transistors. Unsurprisingly, Bardeen soon quit Bell Labs to take an academic post in Illinois. He was so disgusted, in fact, that he gave up semiconductor research.
Things turned sour for germanium, too. By 1954, the transistor industry had mushroomed. The processing power of computers had expanded by orders of magnitude, and whole new product lines, such as pocket radios, had sprung up. But throughout the boom, engineers kept ogling silicon. Partly they did so because germanium was temperamental. As a corollary of conducting electricity so well, it generated unwanted heat, too, causing germanium transistors to stall at high temperatures. More important, silicon, the main component of sand, was perhaps even cheaper than proverbial dirt. Scientists were still faithful to germanium, but they were spending an awful lot of time fantasizing about silicon.
Suddenly, at a semiconductor trade meeting that year, a cheeky engineer from Texas got up after a gloomy speech about the unfeasibility of silicon transistors and announced that he actually had one in his pocket. Would the crowd like a demonstration? This P. T. Barnum—whose real name was Gordon Teal—then hooked up a germanium-run record player to external speakers and, rather medievally, lowered the player’s innards into a vat of boiling oil. As expected, it choked and died. After fishing the innards out, Teal popped out the germanium transistor and rewired the record player with his silicon one. Once again, he plopped it into the oil. The band played on. By the time the stampede of salesmen reached the pay phones at the back of the convention hall, germanium had been dumped.
Luckily for Bardeen, his part of the story ended happily, if clumsily. His work with germanium semiconductors proved so important that he, Brattain, and, sigh, Shockley all won the Nobel Prize in physics in 1956. Bardeen heard the news on his radio (by then probably silicon-run) while frying breakfast one morning. Flustered, he knocked the family’s scrambled eggs onto the floor. It was not his last Nobel-related gaffe. Days before the prize ceremony in Sweden, he washed his formal white bow tie and vest with some colored laundry and stained them green, just as one of his undergrad students might have. And on the day of the ceremony, he and Brattain got so wound up about meeting Sweden’s King Gustav I that they chugged quinine to calm their stomachs. It probably didn’t help when Gustav chastened Bardeen for making his sons stay in class back at Harvard (Bardeen was afraid they might miss a test) instead of coming to Sweden with him. At this rebuke, Bardeen tepidly joked that, ha, ha, he’d bring them along the next time he won the Nobel Prize.
Gaffes aside, the ceremony marked a high point for semiconductors, but a brief one. The Swedish Academy of Sciences, which hands out the Nobel Prizes in Chemistry and Physics, tended at the time to honor pure research over engineering, and the win for the transistor was an uncommon acknowledgment of applied science. Nevertheless, by 1958 the transistor industry faced another crisis. And with Bardeen out of the field, the door stood open for another hero.
Although he probably had to stoop (he stood six feet six), Jack Kilby soon walked through it. A slow-talking Kansan with a leathery face, Kilby had spent a decade in the high-tech boondocks (Milwaukee) before landing a job at Texas Instruments (TI) in 1958. Though trained in electrical engineering, Kilby was hired to solve a computer hardware problem known as the tyranny of numbers. Basically, though cheap silicon transistors worked okay, fancy computer circuits required scores of them. That meant companies like TI had to employ whole hangars of low-paid, mostly female technicians who did nothing all day but crouch over microscopes, swearing and sweating in hazmat su
its, as they soldered silicon bits together. In addition to being expensive, this process was inefficient. In every circuit, one of those frail wires inevitably broke or worked loose, and the whole circuit died. Yet engineers couldn’t get around the need for so many transistors: the tyranny of numbers.
Kilby arrived at TI during a sweltering June. As a new employee he had no vacation time, so when the cast of thousands cleared out for mandatory vacations in July, he was left alone at his bench. The relief of silence no doubt convinced him that employing thousands of people to wire transistors together was asinine, and the absence of supervisors gave him free time to pursue a new idea he called an integrated circuit. Silicon transistors weren’t the only parts of a circuit that had to be hand-wired. Carbon resistors and porcelain capacitors also had to be spaghettied together with copper wire. Kilby scrapped that separate-element setup and instead carved everything—all the resistors, transistors, and capacitors—from one firm block of semiconductor. It was a smashing idea—the difference, structurally and artistically, between sculpting a statue from one block of marble and carving each limb separately, then trying to fit the statue together with wire. Not trusting the purity of silicon to make the resistors and capacitors, he turned to germanium for his prototype.