Ralph McNutt’s recent update on the progress of the Innovative Interstellar Explorer concept elicited plenty of comments, enough that Dr. McNutt wanted to answer them in a new post. Now at Johns Hopkins University Applied Physics Laboratory, McNutt is Project Scientist and a Co-Investigator on NASA’s MESSENGER mission to Mercury, Co-Investigator on NASA’s Solar Probe Plus mission to the solar corona, Principal Investigator on the PEPSSI investigation on the New Horizons mission to Pluto, a Co-Investigator for the Voyager PLS and LECP instruments, and a Member of the Ion Neutral Mass Spectrometer Team on the Cassini Orbiter spacecraft. With all that on his plate, it’s hard to see how he has time for anything else, but McNutt also continues his work as a consultant on the Project Icarus interstellar design study. His Innovative Interstellar Explorer is a precursor mission designed to push our technologies hard.
by Ralph McNutt
I typically do not get involved with commenting on comments just because of the time constraints of protracted discussions, but some of the questions raised by your readers are, I think, very good and deserving of a response. [The original post is Update on Innovative Interstellar Explorer — readers may want to skim through the comments there to get up to speed — PG].
Let me try to take the comments, for the most part, in order. At one point we did take a look at Sedna and the other large trans-Neptunian Objects (TNOs). The orbit of Sedna (can be found here) will move through ~60° of arc and through its perihelion between now and 2100 (just prior to the aphelion of Pluto) — this is out of an orbital period of ~12,600 years. All of this motion is within 90 AU of the Sun, the orbital inclination is ~12° and is certainly accessible with the appropriate “tweak” at a Jupiter gravity assist. Such an aim point also puts constraints on exactly where with respect to the direction of the incoming interstellar wind one is aiming. To exit the solar system rapidly, one wants a speed as high as possible. Traveling “only” ~17 km/s (about the flyby speed of Voyager 1 past Titan and faster than the speed of New Horizons past Pluto of ~13 km/s), close imaging is problematic (with a radius of 1500 km, this is an object radius travelled in ~100s). Several months of high resolution imaging are possible with a large camera such as LORRI on New Horizons but not with a cell phone camera (which would die rapidly in the space radiation environment shortly after launch anyway).
Eris, with an orbital inclination approaching 45° is currently about 97 AU from the Sun (orbit here) and is inbound to perihelion, crossing the plane of the ecliptic at ~90 AU in the early 2070’s, but still outside of 83 AU in 2100. Makemake (orbit) is currently well above the plane of the ecliptic, passing through the plane of the ecliptic just after the end of this century and just inside of 50 AU; its orbit has a relatively small eccentricity of ~0.16 and an inclination of ~29, etc. The real problem is that doing a flyby of a TNO really is a different mission.
It is perhaps also worth noting that nuclear electric propulsion has been looked at – and in some detail under NASA’s Project Prometheus. The problem is that the power system needs to have a specific mass no greater than ~30 kg/kW (something noted by Ernst Stuhlinger back in the 1960’s — Stuhlinger literally wrote the book on ion propulsion) to have an advantage in speed delivered by nuclear electric propulsion (NEP). But that has to include the mass of the system for dumping the waste heat of the reactor (from the second law of thermodynamics) as well as its mechanical supports. The Prometheus architecture came in at over twice that, and that is the problem. To date all NEP designs come in underpowered when engineering closure on the system as a whole is examined. Think of Hiram Maxim’s steam-powered airplane versus the gasoline-powered airplane of the Wright Brothers. This is ultimately the problem with VASIMIR as well – a more mass-efficient means of providing the wall-plug electricity is needed, if it is to ever become a real system.
The spacecraft mass question is a good one as well. We tried pushing that on the precursor to IIE that was funded by NIAC – an “all the stops pulled out” approach that reduces the spacecraft mass to ~150 kg including a payload. Again the problem is engineering closure. Even if I miniaturize the electronics to microminiaturized solid state items, I need communications, guidance and control, power, thermal control, and a payload. The payload sensors have to be a finite size just to collect the data if all I am fighting is Poisson statistics – which can be traded against integration time (but it makes no sense to spend 10 years to make one measurement). Even with an iPad or equivalent that is radiation hardened, one cannot reduce the mass arbitrarily and then still make the measurements that are the raison d’etre for the effort in the first place. Ultimately, one runs into physical limits set by the properties of the materials from which one constructs components.
Part of this is manufacturing and part is the physics of the material itself. Practicalities are also involved. For the NIAC effort, we looked at the idea (and not a new one) of using ultra-low power (ULP) electronics running at liquid nitrogen temperatures. But now I have a real problem in testing such devices, as the coefficients of thermal expansion of the materials as well as the Johnson noise can preclude operation at room temperature. I could fix that with a lab and facilities on the Moon, but now that infrastructure is required, and the technicians would have to work in space suits – and I have a scenario that does not close economically (and may not technically either). Everyone in the deep-space robotic business has mass reduction as a primary goal – on everything. One can build *something* for less than ~250 kg, but the indications are that to build the desired functionality, that type of mass limit will be “sporty.”
Image: IIE Initial Concept Closeup. Credit: JHU/APL.
With respect to launch vehicles, the use of “really, big” vehicles for robotic missions has always been problematic because of the cost. There was a Voyager Mars Program in the late 1960’s which envisioned using a Saturn V to send to large rovers to Mars. Similarly we looked at implementing IIE with an Ares V combined with either a Centaur or NERVA upper stage. While flyout times are reduced, the decrease is not a factor of two.
With respect to communications, in the NIAC work we looked at an IR optical communications systems running at about 890 nm (see this paper). That was not the problem. The problem was holding spacecraft pointing well enough to keep the laser spot on the Earth from 1000 AU (the requirement for that more aggressive mission). One can certainly do the pointing with a sufficiently capable guidance system – but that drove the mass even more. We found that the best trade was a high gain antenna (HGA) of just under 3 meters diameter (about what is on Pioneer 10 and 11 and New Horizons). One driver is holding tolerances during manufacture and another is holding them under the vibration environment imposed by the launch. Materials are not infinitely stiff (which is good, because then they would break), but that means corrections and feedbacks as required. The other interesting thing about a laser com system running from ~5 light days out is that closed-loop operation is not credible, and the beam is sufficiently small and the distance sufficiently large that to minimize power, I need a clock with an ephemeris that can be used to point the transmitter to where the Earth will be when the modulated laser carry arrives there.
Ralph,
Thanks so much for these details, as it helps connect the ambitions to the realities of the technology.
~Marc
Yes, I absolutely must echo Marc Millis’ comment. The information on NEP especially was enlightening to me. I’m somewhat familiar with the mass issues of ‘hardening’ modern computer equipment, but I had not previously read that Prometheus was as far as a factor of 2x off from mass viability.
We have some work for our next generation of engineers to do, if we can inspire them and train them decently.
So fission reactors are out, and Tokomaks are out, but Polywell and Deep Focus are maybe’s (if they work, of course). Most especially if the p-Boron11 fuel cycle is usable.
Ralph, in your last part with respect to communications, wouldn’t a interplanetary network of satellites (Internet in space) be the best option to eliminate the large hi-gain antennae and cut down on power requirements as well as weight?
This is a great summation of the trade-offs that go into the design of just a precursor mission. Image what the trade-off list must be like for a true interstellar mission. I recall reading once that the Apollo missions required that something on the order of ~10,000 questions needed to be addressed. How many questions must we ask before starflight is possible?
This is real engineering, but the sense I get is that a lot of science, new and exciting, needs to be done before even the IIS leaves earth.
I’d like to echo the praise of this article. The details are important and appreciated.
‘One can build *something* for less than ~250 kg, but the indications are that to build the desired functionality, that type of mass limit will be “sporty.”’
Then instead of asking, “what’s the minimum mass for the desired functionality,” and coming up with an answer that’s higher than politicians are willing to pay, we should be asking: what functionality can we desire that is affordable? For example, what functionality can we desire that can be fitted onto a spacecraft that is 25 kg instead of 250 kg? How about 2.5 kg? There are a very wide variety of kinds of sensors and surely there are at least a few varieties that can be sufficiently miniaturized (or already miniaturized versions already used on earth that can be sufficiently hardened) to meet the constraint.
The main political selling point isn’t even sensors or science in the first place. It’s being the first to go somewhere where no country has gone before. Sputnik was politically spectacular and had no sensors at all.
Thank you for a most complete and enlightening explanation of the engineering issues.
Two observations:
(1) “that has to include the mass of the system for dumping the waste heat of the reactor (from the second law of thermodynamics) as well as its mechanical supports. The Prometheus architecture came in at over twice that, and that is the problem.” — in other words, the cost of converting energy from nuclear fission to to electricity is too high, most obviously because the inefficiency of the conversion has to be dumped as waste heat, which requires massive radiators. Less obviously, but even more importantly, the energy waste is too high because of the high exhaust velocity of ion engines — the power required (and thus for a given efficiency the radiator mass) is proportional to the square of the exhaust velocity.
(2) If propellant is very cheap, the other end of the rocket equation starts to look attractive. Although propellant mass consumed must increase exponentially, the amount of energy and radiator mass required decreases as the square. As a result, when the constants are sufficiently divergent (the cost of propellant mass is sufficiently cheaper than the cost of energy) and the target delta-v is not too high it makes sense to lower exhaust velocity rather than boost it.
Enter the Big Dumb Interstellar Explorer. The key idea is to mine the propellant from an asteroid and store it as a solid. No propellant _or_ tank launched expensively out of earth’s gravity well required. Water can be be stored as a solid outside the “ice line” without a sunshade. Substitute thermal rockets for electric (ion) rockets, a much more efficient conversion of energy. This, and even more importantly the lower exhaust velocity, reduces the required radiator mass to insignificance. Assume we can buy a miniscule fraction of the the water on Ceres for 10 cents per kilogram (obviously it makes no economic sense to mine Ceres just for propellant for such a mission, so I’m assuming these costs are shared with many other applications). This is in the vein of the futurism of previous posts, not in terms of basic technological or scientific advance (the technology used here is actually more crude and primitive than proposed in the OP), but simply in assuming such a mining business.
Under such assumptions we can achieve the following:
Final velocity of 25 km/s without Jupiter gravity assist (unfortunately we are necessarily limited to not go much above this velocity by using such a low exhaust velocity propellant: increasingly final velocity by 3 km/s requires increasing propellant mass by a factor of 10).
Payload mass of 5,000 kg (20 times as much as needed for the “minimum desired functionality”!)
Total cost: $1.8 billion
Details:
delta-v, m/s 24858
thermal engine exhaust velocity, m/s 2000
full mass, kg 2.50E+09
empty mass, kg 1.00E+04
payload mass, kg 5000
total energy involved in accelerating propellant to exhaust velocity, kg*(m/s)^2 (joules)
9.99996E+15
energy density U238, joules/kg 2.00E+13
energy efficiency of thermal rocket 50%
implies mass of U238 needed 1.00E+03
suggests mass of nuclear reactor needed 5.00E+03
cost, dollars, of nuclear reactor and thermal rocket, and transporting these to asteroid belt at $40,000/kg
2.00E+08
cost, dollars, of propellant at $.10/kg
2.50E+08
cost of instrument payload 1.35E+09
total cost 1.80E+09
cubic km water estimated on Ceres 2.00E+08
kg water estimate on Ceres 8E+39
propellant mass as fraction of water on Ceres 3.13E-31
@Joy. I think you are unduly pessimistic assuming just 2km/s exhaust velocity.
If the water was electrolyzed at Ceres, you could have a regular bi-propellant engine with exhaust velocity in excess of 4 km/s.
I’d be looking for a mass tradeoff to see whether microwave or similar electrothermal engines with even higher exhaust velocities (double?) made sense.
Alex Tolley: “If the water was electrolyzed at Ceres, you could have a regular bi-propellant engine with exhaust velocity in excess of 4 km/s. ”
But then you’d have to store the propellants in tanks, which would be orders of magnitude heavier than the nuclear reactor. That means orders of magnitude more non-payload empty mass in the rocket equation, and orders of magnitude more mass that would have to be launched from earth. It’d economically ruin the project.
Storing the propellant as a solid is a huge economic breakthrough.
Possibly one could store the water as ice and electrolyze it only just before it is used en route with nuclear electric power, but this would reintroduce the thermal inefficiencies (requiring high radiator mass) of nuclear electric power discussed in the original post, albeit at lower power levels. It would be worthwhile to do a tradeoff study of these two alternatives (thermal vs. bipropellant with last-minute electrolysis), I imagine.
One way to sharply reduce the mass of reactors is to increase the temperature at which heat is radiated (T^-4, according to Stefan-Boltzmann). I have often thought about a “lightbulb reactor”, a solid state configuration of highly refractory fuel and moderators that would use direct radiative cooling. Energy conversion would be thermo-electric, which can be very mass efficient. The whole thing would operate at temperatures above 2000K, at the limit at which structural integrity could be maintained. Waste heat would be radiated directly, without fluidic or active systems of any kind (although heat pipes might be considered also). Carnot efficiency will not be great at such high rejection temperatures, but this should be more than offset by the dramatic reduction in radiator and overall system mass. I find almost nothing like that in the literature, perhaps it is not feasible for some reason?
Absent a “lightbulb reactor” plus ion drive, I think Nick is right that solid ice as fuel for a nuclear thermal rocket could be the next most attractive solution. It would not offer great exhaust velocity, but it avoids the heat rejection problem and the lack in Isp can be somewhat (but not much) offset by the much larger mass ratios achievable. Solid fuel is also a good option for interstellar travel, although ice would not do in this case. It would have to be Uranium, Plutonium, Lithium deuteride, or any other pure, solid nuclear fuel.
Ion engines, or so I understand, require relatively high exhaust velocities for efficiency’s sake, because of the energy cost of ionizing the propellant. In theory, though, you could run something similar to an ion engine, using dust, and efficiently achieve exhaust velocities intermediate between chemical propulsion and true ion propulsion, because the mass to charge ratio would be much higher than for an ion. That might hit a sweet spot.
@Nick
Storing the propellant as a solid is a huge economic breakthrough.
I agree. But let’s be careful about handwavium here. Our ice ship would have no propellant tanks, but it would need equipment to extract small amounts of the ice and feed it to the thermal engine. That is probably going to be electrically powered, and so you are back to the nuclear reactor thermal inefficiencies again, although one might hope to use the ionized exhaust to generate that electrical power.
Where water ice really stands out is for manned flight. The propellant is ubiquitous. It can be used in multiple ways – meteor shield, radiation shield, consumable water, bathing, emergency O2 via electrolysis, converted to H2O2 for simple engines, etc.
Eniac, your “lightbulb” reactor sounds like a great idea, but one that alas may fall through the cracks, because “obviously” one is supposed to try to maximize the Carnot efficiency (“obviously” on earth of course, because cooling is so easy down here, but here is where the nuclear engineers learned their trade). Similarly working against Zuppero’s idea of nuclear thermal “steam” engines using native propellants (and my related idea of storing such propellant as a solid until used) is the dominant belief that high exhaust velocity is always better, because launching propellants and tanks into space has always been expensive.
Up to about 20-25 km/s with very cheap and containerlessly stored propellant the low exhaust velocity thermal engine is in fact the most economically attractive, but I should point out that if we want to do a pure interstellar mission (i.e. no slow flyby of a Kuiper or Oort object), the low exhaust velocity does indeed start to impose a huge barrier to higher final velocities. It doesn’t make sense to try to push it much above about 25 km/s. For higher final velocities, just-in-time bipropellant, or if feasible Brett Beltmore’s suggestion of a dust ion engine (if the dust can be mined and used from the moon or asteroids, stored with little or no container via electrostatic attraction, and then used in a rocket engine, all without too much heavy processing equipment brought from earth) will work better. If we don’t mind taking longer (even hundreds of years?) to achieve final velocity the radiator problem (proportional to power rather than total energy) diminishes, making the dust ion option look even more attractive. Eniac’s “lightbulb reactor” also makes the native propellant ion idea even more attractive.
For long-term interstellar design, high exhaust velocity _and_ containerless storage of native (mind from space rather than earth) propellants are both very important, as are designs (if electrical power is used) that optimize the entire thermal process rather than just the Carnot efficiency. But alas this a very little explored territory even in interstellar spacecraft design circles, because of earthbound assumptions made about thermal design, propellant expense, and manufacturing capabilities in space (the latter assumptions being either none at all or just like those on earth — both will prove to be profoundly wrong).
Alex, once can use a small fraction of the thermal energy itself to melt the ice, and with a bit of cleverness thermal energy can also power the pump(s) and injectors without the intermediation of electricity. One would only introduce electrical energy for this equipment where it makes the task simpler and more reliable without adding substantial mass. Since we are still talking about a very low thrust rocket, the power levels and mass of equipment involved to melt the ice and pump the resulting liquid to the engine are very small.
I feel that something is being missed (possibly by me) in talk of using free chunks of asteroid to produce most of the bulk of a space ship propellant. If we can do that, why can’t we instead use an outer icy moon of Jupiter. Thereafter at launch we can quickly utilise the slingshot effect of the Galileans to boost the probes speed.
And here is where I have a problem. We can also use the slingshot effect of the Galileans to cheaply bring that rocket very close to the Jovian surface, whereupon the Oberth effect would offer such high potential for boosting speed that it would become tantalising to redesign that rocket so that can produce a high delta V in short space of time. Surely that is cheaper still for producing solar hyperbolic excesses of 20 km/s.
I should have put in some figures to quantify my above problem. If I wanted a Jovian hyperbolic excess of 20 km/s, I could achieve half of it if I could produce a high thrust delta V of just 900m/s and all of it if a high thrust delta V of 3.5km/s can be produced.