Click to viewWelcome back to Ask a Biogeek, a biweekly column where UC Berkeley biology researcher Terry Johnson answers your questions, no matter how weird. Reader Mike asks:

Can you speculate on what a silicon based lifeform might look like? What would an "organic chemistry" look like for silicon, instead of carbon?

Life on earth is (so far as we know) exclusively carbon-based. Thanks to its position on the periodic table, carbon can comfortably form bonds with up to four other elements - including other carbon atoms - allowing it to form a wide variety of complicated molecules necessary for terrestrial life. Silicon, being right below carbon on the periodic table, is chemically similar in many ways, leading science fiction authors to consider the possibility of life with a biochemistry that switches out carbon in favor of silicon.

In 1894 H. G. Wells wrote:


One is startled towards fantastic imaginings by such a suggestion: visions of silicon-aluminium organisms – why not silicon-aluminium men at once? – wandering through an atmosphere of gaseous sulphur, let us say, by the shores of a sea of liquid iron some thousand degrees or so above the temperature of a blast furnace.

Since then writers have imagined silicon-based creatures as diverse as Star Trek's Horta and the Xenomorph (though I may be cheating here, since it's unclear how rigidly the xenomorph adheres to silicon-only biochemistry).

"I'm a doctor, not a bricklayer!" - a clear example of carbon chauvinism.


Silicon is also a major component in microchips, so one can make a case that an artificial intelligence would be a silicon-based lifeform. So, which is more likely - stumbling upon silicon-based biochemistry out there amongst the stars, or creating life that thinks with silicon-based microchips here on earth?

Silicon is the most abundant element (barring oxygen) in the earth's crust. If silicon is so chemically similar to carbon and it's so readily available, why aren't we silicon-based? Silicon is routinely used by carbon-based lifeforms, but while (for example) diatoms (a type of algae) make their cell walls out of silica, carbon in is the backbone of their DNA, their proteins, and the basis of their biochemistry. Silicon is just along for the ride.

The answer involves subtle differences between carbon and silicon chemistry. While carbon and silicon can theoretically form very similar kinds of structures, complicated carbon-based molecules tend to the stable, while complicated silicon-based molecules tend to fall apart (especially in water).

There's a major waste disposal issue as well - carbon dioxide is a gas, and silicon dioxide (sand) is a solid. When we metabolize oxygen, we produce carbon dioxide as a waste product, but it dissolves easily in our blood for rapid waste management. If, on the other hand, we produced sand internally with every breath, chaffing would be the least of our worries. Most silicon molecules also lack chirality (or "handedness"), which is a hallmark of terrestrial carbon-based life, but not necessarily a deal-breaker.

I won't go so far as to say that there's no such thing as a silicon-based biochemistry. As Arthur C. Clarke said, "When [a distinguished and elderly] states that something is impossible, he is very probably wrong." Being relatively young and almost completely undistinguished, my odds would be even worse. I will say that, if a silicon-based biochemistry exists, it probably doesn't use silicon the way we use carbon, and we might even have a difficult time recognizing it as life (unless it mind-melds with Spock).

Though silicon would be the basis for a chancy biochemistry, it makes (in part) a fine integrated circuit. As computation and storage become less expensive, our knowledge of how living things think has expanded, thanks mostly to increasingly powerful experimental techniques. Our brains are Gordian Knots of neurons; a tangle of cunningly interconnected cells from which consciousness arises. If we'd like to replicate that consciousness in silico, we need to do more than untie the Gordian knot - we need to somehow ascertain which strands of the intact knot interact to understand and reproduce the brain's wiring.

Alexander the Great's solution is pretty close to the mark, though he'd have needed a thinner sword. First, you section the brain into microscopically thin slices, then you image the slices. Reconstruct the images into a 3D model and you might be able to tease out which neurons communicated with each other.

A fruit fly brain - an image that took 10 months to produce.

It's also possible to help distinguish individual neurons by coaxing neurons to fluoresce different colors, a technique aptly called a brainbow.

The amazing technicolor brainbow.

Once you have a wiring diagram for the neurons in a brain, you can run a computational model of how they'd interact. Today it takes a supercomputer like Blue Brain to simulate even part of a rat's brain, but that part (the neocortal column) looks to be reacting much like a real rat's would.

A virtual rat neocortal column, according to Blue Brain.

I expect that the first silicon-based life will be a simulation of carbon-based life. While our brains rely on their three dimensional structure, ion channels, and neurotransmitters to do the grunt work of consciousness, a simulated brain could achieve the same end result without the benefit of carbon. With a little luck these exacting imaging studies will become become possible without having to chop a brain into thousands of thin slices.

When that day arrives, you'll see me stocking up on hard drives.

Want to ask Terry a question? Email him!