Tuesday, January 18, 2011

Mining Asteroids is Hard

With the costs of rare earth metals on the rise, why can’t space entrepreneurs mine asteroids for platinum and other REM’s and return the materials to earth? Shouldn’t finding so many near earth asteroids make the problem even easier to solve (less delta-v to reach these nearby asteroids)?

Usually this blog focuses on the positive – on the how you could make this happen. Today we are going to look at how hard it actually would be to close such a business case.

  • Mission: Mine platinum on NEOs and return the processed ore to earth for sale and consumption. Sale of platinum sole revenue source for the mission.
  • Mining Efficiency: for every one kilogram of mining equipment launched, the machinery could mine 100 times that amount of NEO material (2500kg mining device could mine 250,000kg of NEO material)
  • Mining Device mass: 2500 kg
  • Platinum concentrations on the NEO: 0.3%
  • Price of Platinum per kilogram: $58,500
  • Mission Cost: $600M

Based on these assumptions, the sale of the platinum mined on the asteroid would cover 7% of the mission costs. This business plan stinks. Not 7%, that seems too small. Really? Only 7% of mission costs could be covered with the assumptions above? Well how elastic are these assumptions? How far would we have to modify the assumptions to get more satisfying results?

Below I explored five what-if’s:
  1. What if platinum was found in higher concentrations?
  2. What if the mining device could mine more?
  3. What if the price of platinum were higher?
  4. What if mission costs were reduced?
  5. A Hybrid what-if.
What if platinum was found in higher concentrations.
The table below shows platinum concentrations would have to exceed 4% to cover mission costs.

What if the mining device could mine more.
The table below shows the mining device would need to mine over 1300x its own mass to cover mission costs.

What if the price of platinum were higher.
The table below shows the price of platinum would need to balloon to $800,000 per kg to cover mission costs.

What if mission costs were reduced.
The table below shows mission costs would need to be reduced to $44M.

Baseline Conclusions.
  • Mining asteroids is hard
  • Platinum mining to serve terrestrial applications is ridiculously hard to justify using these baseline assumptions
  • Entrepreneurs may have to seek business plans that fundamentally change these assumptions or offer their product to non-terrestrial customers

A Hybrid what-if.

But I can’t leave a post with such reserved pessimism. The table below shows that if an entrepreneur could find a NEO with platinum concentrations significantly higher than average even while assuming a less efficient mining device, such a mission may be possible if the costs could be reduced to less than $150M.

Have fun (in a nerdy spreadsheet kind of way) building your own platinum mission by using the spreadsheet located here.

[UPDATE: I fixed the spreadsheet so readers can download the file in MS Excel]


  1. Quite nice. We would still need survey missions. Such missions could be acomplished by a specialised using probes or robotic explorers to scan and survey the asteroids in great detail and sell that information to private business. Who knows, maybe we encounter a 15% platinum and other REMs asteroid. Sounds amazing but we dont realy know until we send something there. The upcoming NASA missions to visit a NEO should be very interesting in this reguard.

  2. Doh... Can't edit the spreadsheet...

  3. Michael:

    Try downloading the spreadsheet now.
    it should work for you.


  4. Andre:

    I agree. There may very well be a market for a commercial survey mission (see my NEAP post from earlier this month).

    I think what the analysis in this post shows is that a commercial mining mission to a NEO should be based on a survey data so the team understands the potential profitability of a mining mission.

    And yes, I am optimistic we find some NEO's with more favorable mining conditions.

    Thanks, good points.


  5. The parameter to work on seems to be mission cost. Assuming 2500 kg is the weight of an automated asteroid processing machine (something that chews up rock an spits out ore). Ignore the cost of development for now since we will use the design multiple times. We may be able to launch four of these to LEO on one SpaceX Falcon 9 for $50 million. What we need then is a tug to deliver these to four asteroids and return processable ore at under $28 million per asteriod. That may not be doable with current technology but it sets some new parameters.

  6. https://spreadsheets.google.com/ccc?key=0At1WrSW42PzGdFFIUjcyczVHdkpXTWFhWURfMkU0THc&hl=en

    That's a link to the Google Docs spreadsheet version. I added a tab for looking at bringing the thing back and mining it here rather than taking everything there.

  7. Any how exactly to we land 150-200 thousand pounds of ore on earth to refine, without vaporizing it, killing anyone and without having to recover it off the ocean floor? What cost would that add? Seems to me like that is the killer.

  8. Any how exactly to we land 150-200 thousand pounds of ore on earth

    Attach it to giant mylar trapezoids filled with vacuum in orbit, and then have it float down?

  9. So usually the asteroid mining debate goes something like this:

    Option 1: only bring back the useful ore by mining the asteroid for the useful components. But that is silly, the critic says, why deal with all of that complexity of a remote compact mining device, instead…

    Option 2: bring back all asteroidal material. Mine the ore for useful metals on earth. But that is silly, the critic says, now you have to deorbit massive amounts of asteroidal material (remember baseline platinum levels are 0.3%). How is this cost effective? Instead…

    Option 3: keep the asteroid in orbit, mine the asteroid there and sell its contents for space purposes, like metal trusses for space stations & spacecraft, solar panel components, mass for shielding, etc. This way you avoid ever having to reenter all of that asteroidal material. But that is silly, the critic says, there is no market for the on-orbit products this solution hopes to produce. You have made the solution so complex, it will be prohibitively hard to raise the investment money for such an endeavor, plus the complexity will delay liquidity events to allow for a time-consuming development cycle (space manufacturing center, etc.). Why not develop a compact mining device that can be sent to surface of an asteroid, dig through a bunch of asteroid, find REMs, and just return that few hundred/thousand kilograms to earth? And now we are back to Option 1.

    Did I mention asteroid mining is hard (and the life of a critic is substantially easier). Keep bringing the ideas. The more ideas the better.


  10. Michael:

    Thanks for the spreadsheet update. Let's refine your hybrid option of returning payloads to the surface. I want to think more about mission costs.

    I like your thoughts. Keep them coming.


  11. Crazy Idea:

    Is there any value in using an AUTOMATED scooper tool to return un-mined material to LEO and use a second REMOTELY CONTROLLED device in LEO to mine the material for REMs?

    Maybe not, the delta-v to stop in LEO rather than reenter would be high, but I am trying to reduce complexity. And avoiding “automated element recognition” while on the asteroid so far from earth has got to be a good thing.

    But would the actual mining effort be easier in LEO if a human on earth could oversee the mining process by using tele-operations? And you avoid hauling heavy mining equipment to an asteroid – just have to get it to LEO.

    Mission Architecture:
    Launch one – Scooper device sent to a NEO. Scooper gathers a lot of rock and transports it back to LEO.

    Launch two – Once Scooper successfully returns to LEO, launch a Remote Miner and Return vehicle to LEO. Asteroidal material is mined in LEO. Once mining is complete, REMs are loaded into return vehicle and brought down for terrestrial sale and use.

    This architecture may offer some economies of scale as the LEO mining station may be able to be reused for multiple mining missions. such a mining station might even be serviceable...

    Thoughts? I have not done the math on this, just a crazy idea for tonight.


  12. Yes, I for one agree with the 'automated return to LEO' and mine from there approach. Perhaps actually GTO instead, so as to avoid the accidental orbital decay and allow one 'ground based' facility to control it, to avoiding relaying delays for real time control. Another point in the spread sheet it that you only have one line for precious metals, but presumably there would be several others as well.

    Platinum Price $90k/kg
    Rhodium Price $79k/kg
    Palladium Price $26k/kg
    Gold Price $44k/kg


    I think the sum total for all of these may get you at least half way to 4.1% break-over.

    And now for slightly more 'insane' ideas, could more or less 'lasers' either one of
    a) the US/European Missile defense ('star wars') types, or
    b) Asteroid deflection type (Dr. H. Jay Melosh, solar reflector thing) (OK, that not a laser...but you get the idea)

    Be used to cut pieces off.

  13. Elden,
    I think we're using Platinum as a proxy since its difficult to know the ratios for any given body. I looked and didn't find any data on average ratios either. So I just kept it based on platinum.


  14. Cool, .... another thought on the equipment..


    15k Watt Continuous Laser
    400 lbs
    150k Watt input power (guessing 10% efficient(?))

    Typical industrial laser is around 3kW, good for clean and pretty cuts of 2 Inches of steel, I think.

    Input power .... solar cells...
    1.5kW/m2 * roughly 10% efficient=0.15kW/m2
    so we need 1000 square meters of solar collector!

  15. How would we move the cutter?

    While ion propulsion engines have very little thrust, the rock of concern would have very little gravity. Would it be possible to use ion engines as positioning thrusters? Either as a) make the machine jump to its new location or b) actually drive the laser head by moving the whole thing?

    PS: I changed my mind, I now agree that the tele-operation in LEO would be better; in order to stay inside the Van Allen belt. (The price of electronics goes up about 2 orders of magnitude when you go outside the Van Allen Belt due to radiation hardening).

    1. If you want to do your processing in LEO, that would mean moving huge masses to LEO. Is there a good plan for this? Would the cost of radiation-hardened electronics be significant compared to this?

  16. A Yale Study on the growing terrestrial use of REMs.



  17. You present very good analyses on the profit making potential of space. In regards to your analysis on asteroid mining there is one key fact that can make it more feasible: the amount that can mined for the mass of equipment transported to the asteroid is *much* higher than your estimate of 100 to 1, in fact orders of magnitude higher.
    First, here's an argument posted to sci.space.policy that a low cost lunar lander can be composed from currently existing components when carried to LEO by the upcoming Falcon Heavy:

    Newsgroups: sci.space.policy, sci.astro, sci.physics, sci.space.history,
    From: Robert Clark
    Date: Tue, 31 Jan 2012 10:12:59 -0800 (PST)
    Subject: Re: SpaceX Dragon spacecraft for low cost trips to the Moon.

    Because the delta-V to a NEO is lower than to the Moon, this lander can likewise be used for asteroidal prospecting.
    Now, in regards to the amount that can be mined for the weight of the equipment, note that on Earth mining, the same equipment is used to process the materials mined for years, continually, and at tons at a time. Clearly this amounts to many times more than just 100 times the mass of the equipment used.
    To put some numbers behind this, here is a post to a space forum about lunar mining of gold:

    Mars in a decade?...

    It notes there is portable gold processing equipment capable of processing 300 tons of excavated material a day, at a weight of only 3,000 pounds. This is a ratio of 200 times the weight of the equipment *every day*. Just operating a couple of years would mean the amount of material processed would be more than 1,000 times higher than your 100 to 1 estimate.
    This is for gold, but quite likely analogous numbers would hold for platinum and other valuable metals.

    Bob Clark