Oct 14 2009 3:41pm
Nanotech IS distinguishable from magic

nanobot swarm

After my debut-post disclosure that I write hard SF, you will be unsurprised to know that bad science in stories can annoy me.

It’s not that I’m looking for a Ph.D. dissertation in each story. It’s not that I limit my interests to stories that revolve around known science. What makes me testy—and, on occasion, has had me tossing a book across the room—is absurd deviations from (what should be, anyway) very basic science.

To be clear, my hobby horse is bad science, not the quantity of science, in stories. If a piece of SF illustrates a bit of science without injury to the story, that’s great—but it’s far from necessary. Many a fine SF story uses science or technology merely as backdrop. Many a fine SF story presumes a technological breakthrough and explores its implications without attempting to predict how the thing might actual work. And many an otherwise fine story bogs down in an excessive transfer of background information.

On to nanotechnology. Briefly, for those unfamiliar with the topic, nanotech deals with manufactured objects on the scale of nanometers (trillionth billionths of a meter). That’s smaller than the individual cells in the human body and bigger than atoms. The goal of nanotech is to manufacture things by precisely controlling the placement of individual atoms.

Nanotech is most definitely real—although still emerging—science. In fact, I’ve written a couple of popular-science articles on the subject. Once engineers learn to build with atomic precision, it will mean such neat things as super-strong materials (space elevator, anyone?) and really tiny machines (like a Roto-Router for the arteries).

A common SFnal premise is nanotech-built, nanoscale robots—nanobots—with the ability to make other things.  Including more of themselves. Fictional replicators tend to run amok and remake the world in their image. In a phrase, this is the “grey goo” scenario.  But even when fictional nanobots are better behaved, they show the wondrous ability to produce pretty much anything on demand, and in very short time frames.  

Clarke’s Law posits that any sufficiently advanced technology is indistinguishable from magic. It does not say that a technology we have yet to invent can produce any desired magical outcome. And so we come to my gripe of the day.

One of the bedrock principles of physics is the conservation of energy. In this universe, energy can be neither created nor destroyed. (Why the caveat? Because maybe our universe began in an eruption of energy from a higher-order universe. Cosmology is just weird. Maybe in another post.) Energy can only change forms, as from the kinetic energy of flowing water into electromagnetic energy—electricity—when the flow (which slows down in the process) turns a dynamo at a hydroelectric dam. Matter is a form of energy, too—hence, the energy released by matter-to-energy conversion in nuclear bombs and reactors (including stars).

The biggest fatal flaw in most fictional portrayals of nanotech—what sends those books arcing across the room—is ignoring that the nanobots need energy to do … anything. So how do nanobots get power? From sunlight? Explain why those nanobots will out-compete photosynthesizing algae. Chemicals in the environment? If nanobots forage for their “food,” I see no reason they would be more efficient at it than bacteria. (In both those cases, the living organisms have had myriads of generations to fine-tune and optimize their techniques.) How about beaming energy to the nanobots as radio waves? Matching antenna size to wavelength, it would take X-rays to power nanobots. Powering our nanobots that way wouldn’t be terribly eco-friendly.

A related problem is controlling the bots. Again, radio waves don’t work. Because of their size, nanobots can only transmit or receive X-rays or even shorter wavelength (i.e., higher-energy) signals. Nor is a flood of X-rays an acceptable way to control the Roto-Routers unplugging my arteries! Chemical signaling between nanobots certainly can work—that’s how biology controls activities within and between cells—but messenger molecules take time to get around.

As you can see, I got curious how nanobots, and especially medical nanobots, might really operate. (There are lots more issues, but I’ve run on long enough for one post.) So: I went to a major nanotech conference to meet the experts. I talked to specialists in biology, neurology, biophysics, and medicine. The result is my latest novel, a near-future medical-nanotech thriller, Small Miracles.

If anyone gets struck by a copy of SM thrown across the room, I want to know why. 

Edward M. Lerner worked in high tech for thirty years, as everything from engineer to senior vice president. He writes near-future techno-thrillers, most recently Fools’ Experiments and Small Miracles, and far-future space epics like the Fleet of Worlds series with colleague Larry Niven. Ed blogs regularly at SF and Nonsense.

1. Freelancer
Mr. Lerner,

Kudos on the article. I began my reading life with hard SF, and have always been drawn to such, even though more of my time has been spent in the fantasy/adventure realm in recent years. I too am disappointed when technology is represented in fiction in ways that a modest amount of research would show to be implausible. For that, I will add Small Miracles to my reading list, and we'll see. ;-)

You are exactly correct that energy or information transmission to objects on such a scale must be done in the x-ray or shorter wavelength bands. Even if the bot were able to employ an external strand of molecules as an effective long-wire antenna to detect a longer wavelength, the receiving circuitry still could not demodulate such a signal. So the mini-radar and radio antennas in Fantastic Voyage are out.

Now, a small point of correction. Nano as a prefix references a billionth (1*10^-9). The prefix for a trillionth is pico. It helps to work where laser pulse bandwidths are measured in femtometers (1*10-15).
Edward M. Lerner
2. EdwardMLerner
Nano = billionth. Of course. I claim a synapse misfire.

Now to learn if I can correct an existing post ...

(And thanks for the kind words about the post.)
April Vrugtman
3. dwndrgn
Ha, I even do this with Fantasy when the magic is too improbable to believe or doesn't fit the worldbuilding.

For the most part science is very fascinating to me but completely over my head so it is rare that bad science would stop me reading (because I just wouldn't recognize it) as long as the story itself was good. That being said, I've noticed that most often bad science, bad technology, bad magic and bad writing all point to poor research and development and would also all go together so I wouldn't necessarily have to recognize poor science because I would have already put the book down for poor writing ;-)

And now you see why I am not a writer :-)
James Jones
4. jamesedjones
BIG wake up call for me on nanotechnology. I was introduced to this tech in a terribly soft sci-fi video game called Alpha Centauri. It based part of the game off of the 'grey-goo' idea. So when a friend of mine in our Optics class showed me the article about the folks making pictures with the carbon tubes, the idea that the two were the same technology just never seemed right. After reading your blog, I can see why.

My favorite aspect of Hard Sci-Fi is the 'rules' that apply and keep the story under control. These also helped me to picture the setting as a possibility, or hope for the future (at least as far as the technology). Nanotech always seemed like a rule-breaker to me, so thank you for getting them back into the realm of possibilities.
Nathan Macey
5. nafhan
Very interesting, Edward. I'd be curious to hear your thoughts on the Greg Bear novel "Blood Music". It's been a while since I read it, but I remember thinking it was well done.
As far as sci-fi in general goes. I tend to prefer "hard" sci-fi as well. I consider most TV shows and a lot of fiction to fall into a category I call Science Fantasy, as any pretense of science is pretty much a fantasy.
Edward M. Lerner
6. EdwardMLerner
Hi, Nafhan. It's been a while since I read Blood Music, but I remember enjoying it quite thoroughly.

The biggest difference between Blood Music and Small Miracles is replicators: Greg's story used them and mine doesn't. And the underlying science in Blood Music is biotech rather than nanotech. So perhaps it's not surprising that the two books go in very different directions.
Cory Gross
7. Cory Gross
There ya' go. My own axe is that, contrary to the received wisdom, sufficiently advanced technology is not indistinguishable from magic. The assertion that it is relies on a view of non-European and pre-Modern people as primitive savages and complete dithering morons.

Sorry, that didn't really have much to do with the piece. Like I said, it was an axe I have to grind. Magic looks like magic because of the apparent circumvention of physical laws which technology obeys, no matter how advanced it is.
Cory Gross
8. Andy White
I feel your pain, although I am maybe a little less rigid in my frustrations... Authors merely need to keep it plausible to keep me happy, but its amazing how many authors fail to even do that. I think dwndrgn hit that nail head on.

It also helps if authors/ script writers care enough to give us a fudge that works, a recent example being Defying Gravity (tv show) where they had the characters 1st explain why they appeared to be in a gravity field all the time, and then in a later episode gave further clarification as to why their hair (and other small objects) didn't suffer from the effects of zero G. Yes the explanation might have failed science 101, but it worked to strengthen the illusion of disbelief required for the show to work.
Peter Erwin
9. PeterErwin
It would take X-rays to power nanobots ...

Except you've just mentioned the possibility of powering them via absorption of optical-wavelength photons, so it's not clear to me why you suddenly need X-rays.

Individual hydrogen atoms (sizes much less than 1 nanometer) can absorb and emit 21-cm radio photons, after all. (Not that that would be an efficient means of power transmission, of course!)
Cory Gross
10. pskye
Responding to PeterErwin at #9.

I'm dusting off my eight-years-past physics degree (which I haven't been using much in my current career) to take a stab at this one so I'll defer to Mr. Lerner if he wishes to provide a better explanation.

Essentially there are three completely different processes mentioned in your post: powering nanobots "via absorption of optical-wavelength photons" versus by X-ray wavelengths, and the emission/absorption of photons by hydrogen.

The hydrogen issue is (to me anyway, but I'm a card-carrying nerd) the simplest. Photons are emitted and absorbed here on the scale of an individual atom, heck an individual electron. The electron orbiting a hydrogen proton has a certain amount of energy but not any possible amount. That amount is quantized into specific states and when an electron moves between them it must either lose an amount of energy (equal to the difference between the higher and the lower state) by emitting a photon with that energy or it must absorb a photon of the correct energy to "pump" it up from a lower to a higher state. The "21-cm radio wave" is one of these specific photons resulting from an electron moving from one state down to another (and at this point I'm not even going to get into the ground state and its hyperfine levels, sorry). Anyway, the energy for and from these state transitions are not at all useful for powering nanobots because they are acting at the level of individual orbital electrons (which do happen to be bound and not available for useful currents).

Why X-rays then? For power transfer (and use in signaling as he mentions also) you now have to deal with the "unbound" or "free" electrons in a conductor. This is less a question of the specific energies of photons (as in the case above) and has much more to do with their electric and magnetic fields and here size does matter. Since the physical size of a nanobot (and therefore its power/communication antenna is so small you are limited to wavelengths that can induce current in that antenna. In the most general sense, an antenna can only detect wavelengths twice its own length (I'm pretty sure) and SHORTER. That's the constraint on what part of the spectrum can be used for powering and communicating with nanobots.

Finally, I'm interpreting the "absorption of optical-wavelength photons" to mean something along the lines of solar cells or even as used in a chemical process like photosynthesis (also referenced my Mr. Lerner). Currently solar cells (and other instances of the photoelectric effect) are still inefficient and I actually don't have any idea how they'd be adapted to nanotechnology so I'm happily punting on this one. Personally, I consider photosynthesis (or similar process) to be a more likely power source but I understand I could be way off with that. I don't know a heck of a lot about it since that would have required a whole lot more Chem and Bio than I ever wanted to take.

So that's my take. If you've somehow made your way through that Wall O' Text then congratulations! You win a Kewpie doll! Or something.

I think I hit all the high points but take it with a grain, no, make that a bag of rock salt. I don't remember everything I ever learned in my classes and even what's here isn't everything I do remember. Heck, I had to hit up the Wiki to confirm most of it.
Edward M. Lerner
11. EdwardMLerner
Peter, you raise excellent points.

Optical vs. X-ray: My intent was to discount optical photons as an energy source without recourse to arguments involving physics (like wavelength mismatch). Hence, I only referred to the bots' uphill struggle to compete with evolved light-using entities.

At that point in writing the post, my mind was still on replicators run amok -- and hence on bots in the wild where sunlight might fall. For bots meant to function within the human body, I could have made do with the observation that the body blocks optical wavelengths. We aren't transparent. But the wavelength-mismatch argument also applies: optical photons won't work with nanoscale bots.

Atoms emitting/absorbing energy: You're correct, of course, that atoms and molecules do exhibit this behavior. We now get into quantum strangeness. One photon going through *both* slits of a screen (i.e., producing an interference pattern), unless you have a detector to identify the slit taken -- that 's weird.

(Is it a wave? Is it a particle? It's two, two, two massless objects in one. )

The 21-cm wavelength (microwave) emission of neutral hydrogen is MUCH bigger than an atom, as you say. It takes a long -- and indeterminate -- time (as atoms "view" time) until emission or absorption happens, because the event is so improbable. As you said: not an efficient means of providing power.
Edward M. Lerner
12. EdwardMLerner
Pskye, had I waited a bit longer to start my last post, I see I could have skipped it. Excellent comment.

For nanobots inside the body, of course, photosynthesis won't work because optical light doesn't penetrate. But chemistry still works ...
13. Freelancer
Piggy-backing Psyke @10

The energy levels to which you refer, that an orbiting electron can "jump" between, are named Fermi levels. The acquisition of a quanta of energy raises the Fermi level of said electron, which moves it to an unstable orbit. It cannot sustain this state, and eventually (4-60 nanoseconeds depending upon the type of atom/molecule) gives up the greater energy to return to it's base or stable energy state. The released energy is virtually always in the form of a photon. The wavelength of the photon is dependent upon the atomic signature of the atom from which it was emitted. For example, a Krypton atom delivers photons of a 248nm wavelength, while an Argon atom releases 193nm wavelength photons.

For nanobot powering purposes, neither of these is usable. Being in the deep ultraviolet range, they induce accelerated oxidation of any solid matter they encounter, especially organic matter like human tissue, hence sunburns. And still, the wavelength itself is far larger than a nanobot would be. The third problem is one of bi-refringence and multipath refraction, both of which would distort and attenuate the photon energy to the point of being inefficient as a means of energy transfer.

For Mr. Lerner's comment about photons passing through two screens, Some interesting physics at work today
Cory Gross
14. Rene Arnush
Very interesting article and comments. I always felt, though, that the comment 'tech. indistinguishable from magic' doesn't require "dithering morons" simply no valid reference point. After all, how many people even now know how a phone works? Or a computer? And these people (theoretically) understand the bare bones of the technology. Also, yes, tech follows physical laws...but we don't know all of them, nor how they interact--look at quantum effects.

As far as power--I have a background in chemistry not physics, but I have read some experiments that successfully used Brownian motion to do work. Fermi levels of the electrons were mentioned...but energy is also absorbed/emitted through vibration and rotation of the various molecular groups, a major component in spectroscopy. Whether any of that can be enough to run nanos...*shrug* but it's not as cut-and-dried as all that. In fact, at the nano scale you may deal with serious quantum issues that throw our experience out the window. In most SF I've read that regarding nanos involved errors in preprogramming, which bypasses the whole controlling them issue.

Personally, I don't mind even iffy science(or magic, for that matter) as long as the world is self-consistent.
Edward M. Lerner
15. EdwardMLerner
Rene, thanks for your comments. Harvesting vibrational energy does have potential. The molecular machines in cells rely on it.

(For those unfamiliar with leveraging random motion, here's an excerpt, slightly edited, from a nanotech article I wrote for Analog: Kinesins are motor proteins that pull themselves -- and cargo -- along intracellular tracks called microtubules. The kinesin molecules exploit thermal motion within the cell to take 200-300 "steps" per minute. This is one of nature's neat tricks: using random thermal motion to drive motion in one direction. Symmetry-breaking mechanisms -- effectively like a ratchet and pawl -- favor selected impacts, giving rise to a directional bias.)

Brownian motion absolutely matters at nanoscale. In Small Miracles, it even plays a critical role in the plot. To my knowledge, that may be a first.
Cory Gross
16. Viadd
I still don't see the need for X-rays. An X-ray is a quantum of energy of about a thousand electron volts (eV), whereas molecular bonds are around 1 eV. It is very hard for something on the molecular scale to catch a X-ray and make use of its energy for anything but tearing itself apart.

Chlorophyll molecules are small, but they catch photons that are much larger than they are and convert them to chemical energy. Fluorescent proteins are a few nm, but produce photons with wavelengths, reliably and on cue, through chemical reactions.

If you are working in the wave regime, where you are looking at the collective action of many photons, then you need sizes comparable to the wavelength. But not if you are working in the quantum regime with individual photons.

As for communication to the nanobot, you could have arbitrarily long 'wavelengths' using, e.g. magnetic sensors based on the Zeeman effect. A 10 mT field intensity is a 1, a 20 mT field intensity is a 0, clocked by reversal of field direction. You can clock as slowly as you want, or as fast as it can be reliably sensed, and nobody needs to care that it looks like a complicated EM wave with a wavelength many times as large as a human.
Edward M. Lerner
17. EdwardMLerner
Viadd, very interesting points.

As you say, short-wavelength E-M radiation destroys molecules -- hence, it's not practical for powering nanobots. Capturing individual photons can work -- again, as you say, chlorophyll does it -- for bots in the wild (vs. within the human body). In the wild, the bots must compete for photons with biota that has had billions of years to learn how to most efficiently exploit light.

In some contexts, the low-rate magnetic signaling you describe will work. It works best for immobile bots talking to a central controller. I don't see applying that method to highly mobile bots, bots deep within (say) a human body, or signaling among largely autonomous bots.

The point of my post wasn't to predict *the* way by which a bot must be powered or with which it will be controlled. Depending on the application, the appropriate method(s) may vary. The larger point was that bots need *some* energy source and control mechanism. Reflecting the limits in fiction makes for better and more realistic stories.
Cory Gross
18. Rene Arnush
After I posted, I was musing about the 'gray goo' scenario and it occurred to me that the problem of power assumes treating the nanos as a machine organism. What about chemical energy? At that scale, it may be as appropriate to view them as a molecular device--sort of customized catalyst. If that is the case, all of the energy needed could be obtained from the molecules it is rearranging. In other words, chemical energy vs. physical energy. I suppose it depends a great deal on your definition of 'nanotech'. I would think, though, that larger nanobots, like very tiny robots, could still use a type of chemical power. As far as competing with existing gut instinct is that a nano would be more efficient, not having to deal with some aspects of living organisms; maybe, though, that would only be the larger ones, trees etc, and the single cell organisms would be the major competitors.
Edward M. Lerner
19. EdwardMLerner
Rene, I agree with you: chemical reactions would be -- where the fuel is available -- a viable way to run nanobots. That's how the medical bots in Small Miracles feed themselves.

Medical bots can tap a ready, well-distributed supply of chemicals (e.g., glucose and glycerol in the bloodstream). It's not obvious to me what fuel(s) bots in the wild could scavenge that would be as suitable and widely available.

Re competition between bots and biota: you're right, I see the competition happening between things of similar physical scales. Nanobots would compete with microbes (or, perhaps, cells within a larger critter) that have adapted to use the most readily available suitable chemicals in their environments.

And if a nanobot is powered by the same mechanism as the cells that surround it, said nanobot is, presumably, limited in some way as the cells are. (Analogy: it takes the energy density of gasoline to allow a car to outrun a hay-powered horse.)
Cory Gross
20. Tynan Sylvester
Old post, but why not?

-Nanobots can come together to form emergent higher-order intelligences, possibly upon sensing a need to do so. This could allow them to vastly outcompete bacteria using any number of intelligence-based strategies. Like rapid development of effective disinfectants.

-Nanobots wouldn't have to compete with each other. Coordinated action by many bots could, once again, outcompete bacteria in many ways.

-Nanobots could be designed and dynamically redesigned for any particular purpose by an intelligence (see above). This would allow them to specialize much faster and better than bacteria for local circumstances.

-Lumps of nanobots can, based on their size, form other useful structures. E.g. a lump the size of a golf ball could easily solve the communications problem by simply forming a small antennae the same size as the one in your cellphone. A nano-lump a few feet across could theoretically form a nuclear reactor inside itself, potentially hyper-charging a grey goo growth scenario. If it could get fissionable material, that is.

-Brings me to another point, which is that nanobots would be limited by access to rare elements as much as energy or raw matter.

Overall, let's not put it last the little buggers to figure out ways to outperform their relatively obvious antecedents like bacteria and people. Fictional nanotech often comes with a singularity seeing as they both involve self-replicating possibly self-improving machines. I sense that a future machine superintelligence might find better ways of getting things done using microscopic independent agents than dumb bacteria or even larger gene vehicles like ourselves.
Cory Gross
21. Burnell
Hi, liked your post, one problem I have is related, is that the nanobots in fiction almost always act as if they have matter conversion capability, that is they act as if they can somehow convert all base elements with little energy expenditure into something else. We can do that now, but it requires a particle accelerator and usually makes the object radioactive, but the point is that in fiction nanotechnology never has to do any of that. It would seem logical that that they are only as capable as the amount of energy they use and yet they never are, somehow they always act as if they were powered by antimatter, and at the same time make materials out of nothing, and are relatively easy produce. I mean it took bacteria millions of years to go through all the different types and specializations to adapt to Earth, nanotechnology, even if designed, would have to do the same, simply because it is a machine and must function to the environment it is in. It doesn't just function perfectly in all enviroments.

Anyway old, but good post.

Subscribe to this thread

Receive notification by email when a new comment is added. You must be a registered user to subscribe to threads.
Post a comment