Dyson shells are possible [was Re: Our rocky solar system may be rare]

Robert J. Bradbury (bradbury@www.aeiveos.com)
Mon, 13 Sep 1999 12:31:54 -0700 (PDT)

On Mon, 13 Sep 1999, Robin Hanson wrote:

> Right. Obviously the universe is very big, and we couldn't expect to
> see a single nano-alien half-way to the horizon. But we *can* say things
> about the aggregation of all aliens in a region. For example, aliens
> aren't intercepting more than 1% of the starlight from the nearest 100
> stars, at least if they re-radiate it at obvious IR temps. (For more
> examples like this, see:

I think I did read this paper quite a while ago (the date on the copy I have saved is Sept. 1998). The problem seems to be the 1% argument from Jugaku. I did investigate this a bit more completely and it seems he got it from Papagiannis from "The Search for Extraterrestrial Life: Recent Developments", 1985, pp. 263-270. Jugaku has been doing Dyson "searches" for 15+ years based on the 1% assumption and has looked at probably the number of stars you quote.

The problems with the Papagiannis argument are: (a) He claims based on some simple calculations that you can't build

      a solid Dyson sphere.  These are correct only using his assumptions.
      If you build a light enough Dyson sphere it can be supported by the
      solar wind (with certain navigational problems).  Pohl and Anderson
      figured out how to solve the materials strength problems for a
      heavier Dyson sphere in their book the "Cuckoo".  While the
      requirements are difficult to meet, Robert Freitas feels they
      can be solved using momentum transfer technologies.
      I've also never seen any calculations on variable thickness
      Dyson spheres (as they finally figured out they had to do to
      make Sky Hook cables work).  So real "solid" Dyson spheres are
      still a fairly open question.

(b) Papagiannis did "claim" without any calculations was that:

        "What is possible, however is to have a large number of
         independent space structures in orbit around the star, but
         these would intercept only a relative fraction (~1%) of the
         star's radiation.  Consequently such stars, would display a
         normal spectrum with only a small excess in the infrared."

      This statement only appears valid if:

      (1) you use the material to create planetoids that have gravity; or
      (2) you create O'Neill type colonies out of all of the material.

So, you only get the 1% result if you *assume* that an ETC must construct habitats for pitiful wet-nanotech like us. If you assume they are only going to construct supercomputers in space (that could care less about planetary gravity), then your material gets spread out over a *much* larger area. Robert Freitas has redone my rather simple calculations on the Matrioshka Brain (layered nested shells of orbiting computer platforms) and they do hold up. They can completely enshroud the star. So then the questions become what is the power the star is emitting and how much material do they have? Smaller stars or more material make the power dissipation per m^2 by the outermost layers lower and therefore the temperature as well.

>From a theoretical standpoint (if you live long enough and grow
big enough *or* do stellar mining to reduce the power output of your star *or* build the MBrain around a Gas Giant and use it to fuel thermonuclear reactors) you can make your radiation temperature *very* close to that of the background radiation of the universe (2.7K). This is the most efficient situation from a thermodynamic viewpoint, and therefore a goal advanced civilizations
would presumably strive for. Anders has discussed this somewhat in his Jupiter Brain paper about the tradeoffs between small-hot JBrains and large-cold Jbrains. I believe the natural evolution is from small-hot JBrains *to* large-cold MBrains. Our telescopes are so poor, we *cannot* detect most such objects *even* in our own solar system. I have some articles discussing IR telescopes that say the dust in our system would hide low power or distant sources that have a temperature below 40-60K unless we put the telescope near or beyond the orbit of Mars. Using our *best* telescopes, I believe the best we can currently do is see newly formed (hot) brown dwarfs with temperatures around 800-1000C out to a few hundred light years.

The devil is in the details in these discussions and you have to question the assumptions of people like Jugaku or Papagiannis in light of technologies we can see on the horizon today.

I've been intending to write to Dr. Papagiannis to request his actual calculations, but its not near the top of my todo list yet.