Hard Science, I Don't Understand

[tbeard1999]

...At the nanoscale, the temperature of the machines performing the work will make them melt or boil.

If you reduce the energy to safe levels (temperatures), then the time to complete the task will increase, making conventional manufacturing processes faster. Grey Goo will either replicate like watching a tree grow or it will boil itself away at a visible pace - neither is particularly dangerous.

<golf clap>

A very pithy "game, set and match" on grey goo. Thank you (and Whipsnade).

(I HATE sci-fi type nanotech for the simple reason that it is effectively a Deux Ex Machina capable of *anything*). Most undramatic and most unwelcome in my campaign.

 

[rancke]

The Ancients had solved the problem of grey goo, of course. Good thing no one has found a warehouse full of the stuff.

Yet...


Hans

 

[veltyen]

(I HATE sci-fi type nanotech for the simple reason that it is effectively a Deux Ex Machina capable of *anything*). Most undramatic and most unwelcome in my campaign.

I feel the same way about magic.

*cough* My 3rd level Zhodani casts magic missile at the darkness *cough*

 

[Icosahedron]

Remember that the Thermite paste consumes itself and radiates the excess energy outside the system. The tiny 'nano-welder' needs to supply enough energy to break and reform the bonds in the metal being welded, but the 'nano-welder' must survive the process. Now we are entering the realm of breaking some atomic bonds while not weakening others. That is much harder than simply scaling down the machine.

Let us also examine some basic properties as things get smaller, starting with area and volume.

<snip some good maths>

Note that as the welder gets smaller, the energy density gets larger. Energy density is the same as Temperature, so these smaller 'machines' are getting hotter. That is why microchips have trouble keeping cool as they get smaller. At the nanoscale, the temperature of the machines performing the work will make them melt or boil.

Thanks ATP, that's more food for thought, but not quite game, set and match yet, perhaps.

If I may play my own devil's advocate some more...

Much depends on the amount of energy needed to make/break molecular bonds and whether the energy density required to do that would be sufficient to raise the machine temperatures to prohibitive values. Also, such machines would be shaped for heat dissipation (fractal radiators, or some such), may be immersed in very thermally conductive material and may be paired with other machines whose purpose is heat transport.

The million dollar question for 'thermodynamic impossibility' is whether the bond manipulation process in and of itself requires prohibitive temperatures, and I'm afraid the maths to figure that is, if not beyond my ability, certainly beyond my patience.

If you reduce the energy to safe levels (temperatures), then the time to complete the task will increase, making conventional manufacturing processes faster. Grey Goo will either replicate like watching a tree grow or it will boil itself away at a visible pace - neither is particularly dangerous.

Again, conventional processes may be faster, but they're not necessarily better. eg. Nano-Weld Paste (tm) can weld CrystalIron whilst maintaining the characteristic 'perfect crystal structure' that gives it its strength, whereas normal welding, although faster, will create a weak joint.
The fact that Dissembler goo destroys its target by heat as well as dissembly may not be a concern for its manufacturer.

If nanotech truly works at tree-growth pace, then I'll agree it's dead, but if it's simply gonna take hours instead of minutes, or minutes instead of seconds, the stuff may still be viable - and it's not clear which is the case.

The proof that Drexler nanotech is impossible may be hidden in the maths somewhere, but it's far from obvious in a way that I can hand to a group of argumentative but math-phobic players and say "nanotech is impossible because..." and unfortunately, that's the sort of 'get out of jail free' card I'm hoping you guys could give me.

 

[Whipsnade]

Much depends on the amount of energy needed to make/break molecular bonds and whether the energy density required to do that would be sufficient to raise the machine temperatures to prohibitive values.

Icosahedron,

Okay, imagine that the majority of the hull is made up of titanium atoms so any repair is going to involve making and breaking the bonds titanium atoms. Next imagine what the "nanite" assembler is constructed of...

Also, such machines would be shaped for heat dissipation (fractal radiators, or some such), may be immersed in very thermally conductive material and may be paired with other machines whose purpose is heat transport.

Heat dissipation in the atmosphere is tricky because we're working "at the bottom". Adding a structure big enough to be a "fractal radiator" and you don't have a "nanite" anymore. As for your other suggestions, immersion in other materials or using heat transport machines, what do you think is occurring in the "reactor-based" nano processes I've wrote of earlier?

The million dollar question for 'thermodynamic impossibility' is whether the bond manipulation process in and of itself requires prohibitive temperatures, and I'm afraid the maths to figure that is, if not beyond my ability, certainly beyond my patience.

Prohibitive temperatures in a shirtsleeve environment.

The fact that Dissembler goo destroys its target by heat as well as dissembly may not be a concern for its manufacturer.

Heat isn't a byproduct here. Heat is how you power the nanites.

If nanotech truly works at tree-growth pace...

Again, it works at a biological pace in a "shirtsleeve" environment. Inside a reactor in which you can provide more energy to the process speeds will be faster.

The proof that Drexler nanotech is impossible may be hidden in the maths somewhere, but it's far from obvious in a way that I can hand to a group of argumentative but math-phobic players and say "nanotech is impossible because..." and unfortunately, that's the sort of 'get out of jail free' card I'm hoping you guys could give me.

If they can't understand the heat/power requirments, how about the "fat fingers" issue or the obscene numbers required?

Nanites are molecular sized and any manipulators they have will be molecular sized too. You going to use a molecule to grab a molecule and stick it in a molecule-sized hole; i.e. fat fingers.

The numbers are mind-boggling too. Let's say I've a billion nanites and each of them can pick & place billion carbon atoms per second. It will take me 167.28 HOURS for those nanites to assemble all of 12 GRAMS of carbon. A billion machines performing a billion operations per second and it will still take almost a WEEK to assemble all of 12 grams of carbon. Twelve grams is about the weight of 4 US pennies.

Hope this helps.


Regards,
Bill

 

[Klaus]

This is great stuff! I did my BA dissertation on Drexlerian and Extropian philosophy, and it's nice to have the prejudices that produced in me confirmed by the laws of thermodynamics!

My erstwhile employers at college frequently try to require us to break the laws of physics (fractions of students, teaching 2 classes at the same time on different campuses, etc), but I think such mathematical proofs (even at the layman level they are at) will just get me fired as it would make the management feel stupid (truth hurts).


The best form of 'nanite', which already works at some of these levels, are gene-tailored viruses and bacteria. Tiny robots will never be able to compete with that!

 

[Whipsnade]

The best form of 'nanite', which already works at some of these levels, are gene-tailored viruses and bacteria. Tiny robots will never be able to compete with that!

Klaus,

From what I've seen occurring in industry today, nanotechnology and biotechnology are going to merge.

In a decade or so, both terms are going to refer to what are essentially the same practices and, within a generation, people are going to be wondering why we even had two different terms.

I'd like to caution those reading this thread, that ATPollard, Klaus, and myself are referring to the "gray goo"/"blue paste"/Drexlerian/science fiction fantasy claims surrounding nanotechnology and not the actual uses of and applications for nanotechnology that at occurring right now.


Regards,
Bill

 

[atpollard]

The million dollar question for 'thermodynamic impossibility' is whether the bond manipulation process in and of itself requires prohibitive temperatures, and I'm afraid the maths to figure that is, if not beyond my ability, certainly beyond my patience.

From my earlier analysis, we observed that for geometric reasons, reducing the scale of the object by 1/10 will increase the energy density by 10x. This ratio will hold true for any object that performs work, only the specific values will change. Since all work involves using energy, any useful task will require some energy.

Let us imagine a machine that does something (any task) and accomplishing this task requires a small amount of energy. For a starting point, lets assume a Meter Scale machine (1 meter cube) and a quantity of energy sufficient to heat the machine by 50 degrees Celsius. A fairly modest value for most industrial processes; a machine that runs at 70 degrees C when cooled with normal room temperature air.

A Decimeter Scale Machine (0.1 meters per side = 0.001 cubic meters) would heat itself by 500 degrees C to perform the same task. Hot, but it can be built with reasonable technology.

A Centimeter Scale Machine (0.01 meters per side = 0.000001 cubic meters) would heat itself by 5000 degrees C to perform the same task. Very Hot, but it can be built with state of the art technology.

A Millimeter Scale Machine (0.001 meters per side = 0.000000001 cubic meters) would heat itself by 50,000 degrees C to perform the same task. Too hot for any known materials – perhaps diamonds or other super material.

Jumping to a Micrometer Scale Machine (0.000001 meters per side = 10^-18 cubic meters) would heat itself by 50,000,000 degrees C to perform the same task. Now we are operating at temperatures in the range of the surface of the sun.

Jumping to a Nanometer Scale Machine (0.000000001 meters per side = 10^-27 cubic meters) would heat itself by 50,000,000,000 degrees C to perform the same task. Now we are operating at temperatures beyond comprehension.

Obviously, we can slow down the process to reduce the temperatures. So reducing the 50,000,000,000 degrees C Nanometer Scale Machine to our still impractical 50,000 degrees C would increase the time required to perform the ‘task’ (whatever it was) from 1 day to 1 million days (2,740 years).

 

[Icosahedron]

Trying to deal with several issues by different posters in one reply, so please excuse me if I miss something.

Bill: Yes, it's difficult if the nanites are made of the same material you're manipulating, and you would hope to avoid this whenever possible, but by using inclined planes, levers, compression versus tension, etc. you might still be able to do a surprising amount of work.

Heat is a byproduct of the energy used to power the nanites.

Today's reactor capabilities are tomorrow's shirtsleeve capabilities - unless you 'bottom out' on the fundamental laws of nature - and this is the crux of my argument: I'm not sure that nanite goo will bottom out.

A billion nanites is a very, very small number. The average splodge of goo would probably contain billions of billions at least.


Klaus: I'm not sure we've reached the Laws of Thermodynamics yet, much less had anything proved by them.

Never is a very very long time. One day, tiny machines may far exceed the capability of viruses.

Arthur: The crux of my argument (about nanite goo not bottoming out) is to question whether nanoscale machines would need the equivalent energy to raise a meter scale object by 50 degrees, or whether the energy needed to break enough bonds in enough time is many orders of magnitude smaller than this.

For example, in Nano-Weld paste, the nanites will only have to work around the edges of the paste to stitch molecules together - a 'thin blue line' a nanometre thick, whilst the rest of the volume absorbs and radiates heat.

By comparison, a lump of thermite breaks all the bonds throughout its volume almost simultaneously, and yet if you drop that thermite into a cubic meter of water, you'll barely register the increase in water temperature.

I'm still not convinced that the amount of energy needed to break and make chemical bonds in a reasonable timescale is so great that it will melt the nanites (this is what I understand by thermodynamically impossible). And even if you could prove it to me, I'm not going to be able to prove it to neanderthal players (present players excepted, if you're lurking, guys - I'm talking hypothetical players here). So interesting as the discussion is, it's not really getting me anywhere.

Fat fingers may actually be a more fruitful path, but I don't think that will be a panacea either.

I need to complete the following sentence with cast-iron finality in not more that twenty words:
"there's no grey goo IMTU because..."

and I reckon I'm still stuck with "there's no grey goo IMTU cos I say so."

 

[tbeard1999]

I need to complete the following sentence with cast-iron finality in not more that twenty words:
"there's no grey goo IMTU because..."

"...Traveller's power and tech assumptions can't handle waste heat from grey goo (even allowing grey goo to exist)."

Looks to me like the power issues are fundamental and unless our understanding of thermodynamics change*, grey goo is simply implausible per our current scientific understanding.

*I'd note that a fundamental change in scientific understanding is possible, but is not probable and isn't the same kind of thing as a technologically-driven improvement.

I think that it's been said, but I'll note that natural nanotech handles the heat problem by building stuff really, really slowly. To do the work proposed, it looks like grey goo will produce a staggering amount of heat. So even if it could produce the power necessary to do the required work in a reasonable amount of time -- something else that seems implausible given our current understanding of science -- it cannot effectively radiate the heat in a non-destructive way.

You could also go with the dramatic argument -- "no grey goo because it's a deux ex machina that can do anything...and that's both lame *and* boring..."

 

[atpollard]

Arthur: The crux of my argument (about nanite goo not bottoming out) is to question whether nanoscale machines would need the equivalent energy to raise a meter scale object by 50 degrees, or whether the energy needed to break enough bonds in enough time is many orders of magnitude smaller than this.

Assuming 100% efficiency, the minimum energy required to break a molecular bond is exactly equal to the energy required to raise that substance from its current temperature to its melting point. (That's the definition of melting, the current energy state is high enough to allow the bonds to be broken - as opposed to a solid). The absolute minimum energy required to disassemble a 1 ton block of iron would be equal to the energy required to raise it to the melting point. Assemblers would need to remove that same amount of energy as waste heat as they convert a liquid to a solid.

Try melting carbon.
That's a heck of a lot of energy per molecule times a heck of a lot of molecules.

Slow or Hot are the only choices.

 

[Icosahedron]

I agree slow or hot are the only choices, but it's not clear to me that keeping the machines within their max operating temperature will necessarily result in timescales that are untenably slow.

I can evaporate the 1D6 on my desk with a weapons grade laser in a nanosecond, I can dissolve it in acid in ten seconds. (which is breaking bonds and disposing of the resultant heat without raising itself to carbon-arc temperatures) Is it a problem if grey goo takes thirty seconds to dissemble it?

Right now, on the basis of the arguments raised so far, I have no idea whether it would take thirty seconds or thrirty years to do the job, so from my POV it's very firmly 'case not proven', and personally I'm inclined to think thirty seconds would be nearer the mark.

 

[aramis]

They'll be slower, typically by a couple orders of magnitude, than a robot of human scale doing the same task, simply because it can wick the heat away faster, further, and easier.

Note that very few structures in nature grow more than a cm or so per day. Most don't even hit 1mm per day.

 

[Whipsnade]

A billion nanites is a very, very small number. The average splodge of goo would probably contain billions of billions at least.

Icosahedron,

Okay, let's use my previous example with a billion billion nanites all making a billion operations per second over the same 167.28 hour period. You know have a whopping 19.92 kilograms of carbon after one week's work.

Don't spend it all in one place.

I'm really surprised you didn't blink about the billion operations per second per nanite statistic, that's several orders of magnitude above most proposed operational speeds. Each of the nanites in my example need to "pick & place" a carbon atoms a billion times per second.


Regards,
Bill

 

[atpollard]

I agree slow or hot are the only choices, but it's not clear to me that keeping the machines within their max operating temperature will necessarily result in timescales that are untenably slow.

I can evaporate the 1D6 on my desk with a weapons grade laser in a nanosecond, I can dissolve it in acid in ten seconds. (which is breaking bonds and disposing of the resultant heat without raising itself to carbon-arc temperatures) Is it a problem if grey goo takes thirty seconds to dissemble it?

Right now, on the basis of the arguments raised so far, I have no idea whether it would take thirty seconds or thrirty years to do the job, so from my POV it's very firmly 'case not proven', and personally I'm inclined to think thirty seconds would be nearer the mark.

Lasers and Acid are bad analogies. In the case of the laser, how large was the laser required to evaporate your die? Now start scaling the operating temperature using my first analysis and a nanolaser will operate at sun+ temperatures. Acid stores the energy in a chemical form and consumes itself in the process. Are your nanomachines designed to break one bond and consume themselves in the process?

The speed will be proportional to energy density which will be proportional to temperature. Living processes are the only nanomachines that we know of and they operate in a very narrow temperature range (around 20 deg C). If your man-made nanomachines can operate at 10x the temperature of organic life forms (around 200 deg C), then they will be 10 times as fast as a tree growing or a bone heals or grass decomposes. If your man-made nanomachines can operate at 100x the temperature of organic life forms (around 2000 deg C), then they will be 100 times as fast as a tree growing or a bone heals or grass decomposes.

So far, billions of years of trial and error have not produced any molecular machines that operate much above 100 degrees C, so good luck.

[as a quick aside, the temperatures given are for order of magnitude approximation only. Accurate temperatures involve degrees K, the amount that the energy exceeds ambient temperature and the efficiency of the machine. I sure don't want to get into that level of detail.]

 

[rancke]

I won't dispute that grey goo style nanotechnology is theoretically impossible. The arguments sound reasonable to me and my knowledge of physics is too meager for me to argue against it.

But for purposes of a reasonably-hard-but-not-fanatic SF game, is there any game reason why grey goo shouldn't work? Not in the Third Imperium, of course, because the TL is too low, but why shouldn't the Ancients have figured out a workaround? Something to do with nano-sized subspace heat sinks, perhaps?

Anyway, it's a self-contained impossibility. There aren't any unintended consequences involve. So what if you can dump a canister of goo on a scrapheap and get a spanking new Ancient war machine out of it? What are the odds you can figure out how to control it?


Hans

 

[tbeard1999]

I won't dispute that grey goo style nanotechnology is theoretically impossible. The arguments sound reasonable to me and my knowledge of physics is too meager for me to argue against it.

But for purposes of a reasonably-hard-but-not-fanatic SF game, is there any game reason why grey goo shouldn't work? Not in the Third Imperium, of course, because the TL is too low, but why shouldn't the Ancients have figured out a workaround? Something to do with nano-sized subspace heat sinks, perhaps?

If something is impossible according to current scientific theory, then I would only put it in a Traveller game if it added a significant benefit to the campaign and if it was critical to replicate the setting (FTL drives meet this test).

I can't really see that grey goo meaningfully benefits a campaign. It's much like a ring with a dozen wishes in a D&D game IMHO and just as ill-advised.

 

[far-trader]

It's much like a ring with a dozen wishes in a D&D game IMHO and just as ill-advised.

Probably be pretty safe in our group then

Seems we always wanted a ring of wishes but once we got one it was a long time before we used it (gotta be careful what you wish for, never know when a more important need will crop up, etc.). I think if one of us had dropped a Ring of Infinite Wishes in a game the finders would have never touched it, backed slowly away, and then run screaming from the imagined ref twists behind such an obvious trap.

Ditto for grey-goo I'd think...

"You found* it, YOU try it!"

"I'm not trying it! What if it tries to "fix" me?!"

"Well there's only a little and we don't know how much it will fix. Let's save it for a more important repair, we can fix this the old fashioned way."

* the only way I see grey-goo existing in my game would be an Ancient discovery, some limited (unknown quantity) applicator method, with no instructions and unknown capabilities...

 

[sabredog]

Ditto for grey-goo I'd think...

"You found* it, YOU try it!"

"I'm not trying it! What if it tries to "fix" me?!"

"Well there's only a little and we don't know how much it will fix. Let's save it for a more important repair, we can fix this the old fashioned way."

* the only way I see grey-goo existing in my game would be an Ancient discovery, some limited (unknown quantity) applicator method, with no instructions and unknown capabilities...

It'd be like handing a tube of cyanoacrylate to a 2 year old. Just imagine all the chaos unleashed because of some Ancient version of superglue. I think I might use this some time - it has potential. I can just imagine how many things can get stuck to a player in inconvenient and amusing ways.

"Hey, what's wrong with the air/raft? It won't move."

"Uh, I was trying to fix the chrome I bent last week 'cuz that stuff worked so great on the hull, and spilled some of 'That Stuff' on the fender and it sort of dripped onto the floor under the raft...."

 

[Icosahedron]

Lasers and Acid are bad analogies.

Probably so.

In the case of the laser, how large was the laser required to evaporate your die? Now start scaling the operating temperature using my first analysis and a nanolaser will operate at sun+ temperatures. Acid stores the energy in a chemical form and consumes itself in the process. Are your nanomachines designed to break one bond and consume themselves in the process?

Not after a single bond, probably, but IIRC Drexleran nanites are expendable and self-replicating.

The speed will be proportional to energy density which will be proportional to temperature. Living processes are the only nanomachines that we know of and they operate in a very narrow temperature range (around 20 deg C). If your man-made nanomachines can operate at 10x the temperature of organic life forms (around 200 deg C), then they will be 10 times as fast as a tree growing or a bone heals or grass decomposes. If your man-made nanomachines can operate at 100x the temperature of organic life forms (around 2000 deg C), then they will be 100 times as fast as a tree growing or a bone heals or grass decomposes.

Now that is a good layman-intelligible argument that I could use - cheers.

If the starting point is accurate.

Trees and bones don't grow at uniform speed (look at seedlings, birth and puberty - and just don't mention my damn dandelions - or roaches, or Ebola virus!). Normal organic growth rates are an optimum based on many factors and could be only a fraction of their theoretical maximum rate, but robots are gonna be designed to work near maximum all the time.

So far, billions of years of trial and error have not produced any molecular machines that operate much above 100 degrees C, so good luck.

Yeah, but so far the machines have been made of hydrocarbons, not metal.

[as a quick aside, the temperatures given are for order of magnitude approximation only. Accurate temperatures involve degrees K, the amount that the energy exceeds ambient temperature and the efficiency of the machine. I sure don't want to get into that level of detail.]

You 'n me both. Too much like hard work - I'm already failing to pick up Bill's billions.