The big accelerator 'round the Moon

[halfelven]

Recently, there was a http://www.reddit.com/r/space/comments/96iwv/the_year_is_2035_we_have_a_permanent_base_on_the/" on Reddit titled "The year is 2035. We have a permanent base on the moon and have sent multiple teams to Mars and back. What's next for NASA?"

There were several interesting responses. I posted http://www.reddit.com/r/space/comments/96iwv/the_year_is_2035_we_have_a_permanent_base_on_the/c0bliol" myself (not something for NASA per se, but somewhat related to the context). I'll try to describe it more clearly here:

It's a few decades in the future, we have a solid presence on the Moon, perhaps we have mastered some self-replicating technologies as well - anyway, it becomes feasible to undertake projects at planetary scale or, at least, at Moon scale. Hence, the idea to build a nice big accelerator girdling the Moon on the equator.

First off, it's near vacuum anyway, so the whole thing does not have to be airtight - it can even be an open structure. It does not have to be continuous. It's enough to place segments here and there, each segment appearing above the horizon for both its neighbors. Each segment is, in fact, a linear accelerator (slightly curved maybe), but the whole system, when all segments round the equator are live, becomes one giant circular accelerator, 11 thousand kilometers long.

Each segment can be self-contained, it can function based on energy from the Sun (energy requirements? what is the surface of solar cells needed?) I assume the coils are superconductive but do not require powerful cooling. (shade? solar-based heat pumps?) Definitely some sort of local energy storage is needed, which gets slowly charged between pulses.

Aiming from an upstream segment to the next segment downstream is going to be tough. These things can be anchored in solid moon rock, but I guess even the Moon vibrates and shakes a little, so some sort of dynamic monitoring and aiming might be required.

Advantages? Many.
- Something on this scale cannot be accomplished on Earth (there's an atmosphere, too many distractions from tectonics, weather, biosphere, etc.).
- The "risk" of popping up a black hole, already tiny anyway, does not matter here - if it does implode the whole thing into a singularity, hey, who cares, it's just the Moon.
- An accelerator this big would take us a whole lot closer to Big Bang physics.

Problems?
- Is the vacuum good enough?
- Interactions between the "free range" particle beam and whatever happens to hit it? (cosmic dust, solar wind and photons)
- The Moon rotates. Is that going to make things difficult for the segments, each trying to accelerate and shoot the relativistic beam into the next?
- Finally, a big accelerator like that will give a new meaning to the expression "hyper-relativistic particle". The segments will have to steer the beam a few seconds of arc or so. I wonder if that will create some problems? (synchrotron radiation, stuff like that)

Thoughts? Comments?

Would be nice if somebody could do a back of the envelope estimation for the maximum energy such a device could pump into the particles. I don't trust myself enough to attempt it. Would it get close to the highest-energy cosmic rays? The GZK limit? Close to the "oh-my-god particle"? (50 J) Higher than that?

 

[Vanadium 50]

If you don't think it's important enough for you to do a back-of-the-envelope estimate, why do you think it's important enough for someone else to?

Also, if you are going to postulate technologies that don't exist yet, how can anyone make a sensible estimate?

 

[ZapperZ]

The whole premise here missed completely the major issues and expenses in building a particle accelerator. It has nothing to do with having a "vacuum". That is the LEAST of our problems. It is the size of the RF structures that becomes the accelerating structures. In fact, more than half of the cost of a particle accelerator is in the building of those beasts.

Think of how bad it gets if you have to ship these things to the moon! And I'm not even going to think about bringing up all the superconducting magnets for the steering, etc.

Before you propose one of these things, you should, at least, figure out the main challenges on such a particle accelerators first. What you listed out, really, missed completely all the major issues and problems surrounding the building of one.

Zz.

 

[Bob S]

The moon's circumference, and the length of such an accelerator (invariably a proton synchrotron), is about 10,900 Km. The present CERN LHC (Large Hadron Collider) is about 27 Km, so the ratio is about 400:1. Because the number of components scales roughly as the square root of circumference, this new machine would be roughly 20 times more complex (i.e., 20 times more control points, such as magnets, etc). When considering such things as mean time before failure (MTBF), in order to have any chance of success, the MTBFs of each single component would have to be approaching 10,000 years. How could such a MTBF be verified before construction? Mean time to repair (MTTR) is also an important consideration. Probably all of the components would be built on Earth and shuttled up to the moon. Last September, the CERN LHC failed after 10 days of running. It is now scheduled to restart next November, a 14-month hiatus. In order to have a rapid response team available to repair component failures, teams would probably have to be stationed every 100 or 200 Km or so around the ring (helicopters don't work very well). Repair technicians might have to wear pressurized space suits to repair the machine (CERN technicians already have to wear oxygen masks while helium is in the magnets). I have read that the CERN LHC requires more power, about 120 MW, than the entire city of Geneva. A moon accelerator might require 48,000 MW, the output of 25 2000 MW power plants. Could solar PV cells supply enough power to keep the superducting magnets at 2 kelvin (superfluid helium), and power all the computers and the ramping magnets? The individual LHC experiments now have crews of thousands, including technicians, programmers, and nearly 1000 physicists each. These would also have to be housed in colonies near their experiments. Last of all (and least), the poor graduate students, who are waiting until the machine comes on to do their experiments, will probably get their Ph.D.s about the time they begin collecting Social Security.

 

[flatmaster]

But if halfelven was willing to fund this...?

 

[Vanadium 50]

But if halfelven was willing to fund this...?

With what? This is pure goofballery.

6 trips to the moon cost $150B in today's dollars: $25B each.

An LHC dipole weighs 30 tons and is 14m long. 11,000 km means 7.8 million magnets, or 230 million tons of magnets. Assuming half the weight of the lunar module descent stage could somehow be used as payload, that's 6 tons per trip, or 38 million trips, at a total cost of roughly one quintillion dollars.

That's the Gross World Product for the next 20,000 years.

Goofballery.

 

[hamster143]

We'll have to learn how to build the thing with lunar materials, it'll be much cheaper than to bring thousands of tons of copper wire & such from the Earth.

How much of the mass of an LHC dipole is cryogenics? And how much is necessary because of the need to maintain high-grade vacuum inside the pipe? If I did the math correctly, each dipole contains around 200 kg of copper wire and perhaps half as much of superconducting material.

If we build the accelerator entirely on one side of the moon, and come up with some kind of mirror-like shielding from sunlight during daytime, the whole thing may be operated without any explicit refrigeration.

The temperature achieved probably will not be in single kelvins and, as such, may be too high for the same superconductors used by LHC. But there are two things to consider:

- We can always use high-temp superconductors if we learn to extract needed materials from lunar soil and ore.

- Why do we need superconductors anyway? Because we need extra-strong fields. And why do we need strong fields? Because, roughly speaking, our energy is limited by the curvature radius of the tunnel, times maximum magnetic field, times constant. And, on Earth, we're limited by the dimensions of the tunnel, so we need to maximize the field. On the moon, it may be more cost effective to build a big non-superconducting accelerator with a 500 km radius.

The "risk" of popping up a black hole, already tiny anyway, does not matter here - if it does implode the whole thing into a singularity, hey, who cares, it's just the Moon.

And then we'll have a little benign black hole orbiting Earth, and I bet that more than one physicist would give up his kidney to study and observe the thing :) But general audience will be freaked out at the mere prospect of having a black hole nearby.Overall, I agree, this is beyond current state of technology. I can see this happen 20 or 30 years from now with rise of lunar mining and AI-driven moonwalker robots.

 

[ZapperZ]

We'll have to learn how to build the thing with lunar materials, it'll be much cheaper than to bring thousands of tons of copper wire & such from the Earth.

How much of the mass of an LHC dipole is cryogenics? And how much is necessary because of the need to maintain high-grade vacuum inside the pipe? If I did the math correctly, each dipole contains around 200 kg of copper wire and perhaps half as much of superconducting material.

If we build the accelerator entirely on one side of the moon, and come up with some kind of mirror-like shielding from sunlight during daytime, the whole thing may be operated without any explicit refrigeration.

The temperature achieved probably will not be in single kelvins and, as such, may be too high for the same superconductors used by LHC. But there are two things to consider:

- We can always use high-temp superconductors if we learn to extract needed materials from lunar soil and ore.

- Why do we need superconductors anyway? Because we need extra-strong fields. And why do we need strong fields? Because, roughly speaking, our energy is limited by the curvature radius of the tunnel, times maximum magnetic field, times constant. And, on Earth, we're limited by the dimensions of the tunnel, so we need to maximize the field. On the moon, it may be more cost effective to build a big non-superconducting accelerator with a 500 km radius.



And then we'll have a little benign black hole orbiting Earth, and I bet that more than one physicist would give up his kidney to study and observe the thing :) But general audience will be freaked out at the mere prospect of having a black hole nearby.


Overall, I agree, this is beyond current state of technology. I can see this happen 20 or 30 years from now with rise of lunar mining and AI-driven moonwalker robots.

Again, there is a severe lack of consideration of the accelerating structures, which occupy the BULK of the major expenses at most particle accelerators. People just don't seem to care about the klystrons, the PFNs, the LINACS, and other RF components and structures that make up the "accelerator" part in a particle ACCELERATOR.

The vacuum and the magnet issues, believe it or not, are NOT the major drawback in us achieving the next energy scale in particle accelerator right now, despite the problems the LHC has recently. It is in the acceleration scheme that currently require bigger and bigger accelerators to achieve the 10's or 100's of TeV scale!

Zz.

 

[hamster143]

The vacuum and the magnet issues, believe it or not, are NOT the major drawback in us achieving the next energy scale in particle accelerator right now, despite the problems the LHC has recently. It is in the acceleration scheme that currently require bigger and bigger accelerators to achieve the 10's or 100's of TeV scale!

Like I said, if you want twice the energy, you need twice the field in each magnet, or you need twice the circulation radius. (Which, even abstracting from the need to buy up land and dig another huge tunnel, implies two times more magnets.)

 

[ZapperZ]

Like I said, if you want twice the energy, you need twice the field in each magnet, or you need twice the circulation radius. (Which, even abstracting from the need to buy up land and dig another huge tunnel, implies two times more magnets.)

You're forgetting that you need to have the ability to get TWICE the energy first! The accelerating structures and components typically occupy the significant cost of a particle accelerator.

All of these discussions is really moot, because the OP hasn't come back to respond or rebut any of these.

Zz.

 

[hamster143]

You're forgetting that you need to have the ability to get TWICE the energy first! The accelerating structures and components typically occupy the significant cost of a particle accelerator.

So you're saying that a single RF accelerating chamber, a few dozen meters long, is more expensive than 50,000 tons of magnets and cryogenic equipment?

 

[ZapperZ]

So you're saying that a single RF accelerating chamber, a few dozen meters long, is more expensive than 50,000 tons of magnets and cryogenic equipment?

If you only have one single RF accelerating structure, you won't be needing all those magnets!

We are talking about something the size comparable to the Tevatron, LHC, and the ILC. Do you think they have only one "... single RF accelerating chamber, a few dozen meters long.. "?

Look at the recent report on the design of the ILC.

Zz.

 

[hamster143]

If you only have one single RF accelerating structure, you won't be needing all those magnets!

We are talking about something the size comparable to the Tevatron, LHC, and the ILC. Do you think they have only one "... single RF accelerating chamber, a few dozen meters long.. "?

Yes, I think that the LHC does have a single RF accelerating structure, which accelerates passing particles by ~1 MeV per pass, and it takes around 1 million passes and 20 minutes to accelerate them to 7 TeV. Magnets are needed to keep particles in a circular trajectory.

I don't see how the design of ILC is relevant in this discussion.

 

[Vanadium 50]

Zz, the accelerating structures of the Tevatron (not counting the injector complex) occupy 1% or so of the space. Most of the cost is in magnets, so that the same RF structures can be used many times during acceleration.

There is, however, an important point. The Tevatron takes one minute to ramp from injection energy to collision energy. Using the same technology in this quintillion dollar machine and it would take about a year. You wouldn't get much physics this way.

 

[ZapperZ]

Yes, I think that the LHC does have a single RF accelerating structure, which accelerates passing particles by ~1 MeV per pass, and it takes around 1 million passes and 20 minutes to accelerate them to 7 TeV. Magnets are needed to keep particles in a circular trajectory.

I don't see how the design of ILC is relevant in this discussion.

Zz, the accelerating structures of the Tevatron (not counting the injector complex) occupy 1% or so of the space. Most of the cost is in magnets, so that the same RF structures can be used many times during acceleration.

There is, however, an important point. The Tevatron takes one minute to ramp from injection energy to collision energy. Using the same technology in this quintillion dollar machine and it would take about a year. You wouldn't get much physics this way.

Sorry, I wasn't totally clear of my point. While I said the scale of LHC and Tevatron, I neglected to mention in that post that I was still talking about the next energy scale, which was my point in my earlier post.

The HEP community right now is almost at a stand still as far as the next generation of particle accelerator is concerned. While the pause certainly is due to everyone waiting for the LHC to give some form of a guidance on what energy scale a linear accelerator should be looking at before it gets built, the other reason is that a $10 billion machine is simply no longer something that many countries are willing to fork out easily. Particle collider/accelerator complex will be getting enormously big and beyond what we can afford without substantial changes for many things, including the acceleration mechanism. We are stuck at roughly 40 MV/m gradient on most copper structures before it breaks down. It is why I brought up the ILC because of its size even though it is a single-pass linear collider.

Zz.