Jump to content

Solar Thermal Rocket


Recommended Posts

The asteroid 2015 BZ509 is on a Sun orbit as big as Jupiter but retrograde, elliptic and titled
A recent study suggests that the asteroid is of extrasolar origin, the only matter known in our solar system. So shall we bring samples back from 2015 BZ509? With a reverse orbital speed around 13km/s, the task seems impossible, but the sunheat engine and a flyby at Saturn would enable it, according to my estimates.

A flyby at Jupiter instead lets difficulties arise, but I didn't try hard enough: reach BZ509's orbit, synchronize with the asteroid, wait for favourable positions... A solar sail doesn't look good: at one Sun-Mercury distance it takes 18 years to tilt its orbit by 163°, and at one Sun-Jupiter distance it's not autonomous.


The probe leaves Earth much faster than for a Hohmann transfer, so it arrives at Saturn in only 2.3 years with much speed in a direction mainly outwards, and passes before and above Saturn which deflects it in a reverse orbit in the same plane (tilt 180°). The minimum global speed cost puts the probe in a Hohmann transfer from Saturn to BZ509, but other options would save time. At mid-transfer, the probe corrects the orbital plane by up to 17°, which is costly, and near the asteroid it makes the final push to accompany it. The Earth-to-Saturn leg is much exaggerated in the sketch: displayed too straight, with the Earth at the wrong place.

Reversing the path to come back would need a doubtful combination of positions and cost much speed, so the probe shall head to Earth directly and aerobrake despite the reverse direction. That's about 68km/s, just above the present record, wow. The path is in BZ509's orbital plane, and braking more than Hohmann needs shall let the trajectories cut an other.


The spreadsheet contains my estimates. A few uncertainties remain: I computed for Saturn at perihelion or aphelion only, but this makes little difference and the flyby copes with various angles, so intermediate cases should be fine; and I computed for BZ509 at perihelion or aphelion only,  but we can't wait until Saturn is at the proper place, and for BZ509 I'm not sure that the intermediate cases are as favourable as the extreme ones.

More details and explanations may come some day.

Marc Schaefer, aka Enthalpy

Link to comment
Share on other sites

Here's a sketch of the Saturn flyby to join 2015 BZ509. Probes use to overfly planets forwards to get speed, but this script overflies backwards to obtain the retrograde speed. This loses speed, so the propulsion must provide more, in amounts compatible with the sunheat engine.


Phi is as in the spreadsheet, Psi is the deflection as in Kate Davis' article
The figures are one example from the spreadsheet, where the probe leaves Saturn with no sunward speed for a cheaper but lengthy strict Hohmann transfer to 2015 BZ509. Other cases can be more realistic, other choices better.

Saturn has advantages over Jupiter:

  • The probe arrives with less Eastward speed in the Sun's frame, that's more Westward speed in the planet's frame.
  • Saturn is slower, so the probe loses less speed.
  • The probe gains speed when falling from Saturn to BZ509.
  • Saturn isn't synchronous with BZ509. The launch date and speed adjustments let the probe meet the asteroid, while arriving on time isn't easy with Jupiter.
  • But Saturn moves too slowly to let pick a position, so the probe will improbably arrive at a nodal point of the asteroid's orbit. The correction may demand additional speed which the spreadsheet doesn't include. Combining the correction with the tilt kick or the arrival kick linders the cost.

Mass estimates may come some day.

Marc Schaefer, aka Enthalpy

Link to comment
Share on other sites

Of the huge speed needed to reach the Sun's termination shock at 90AU in bearable time, a slingshot at Jupiter can give 5 to 6km/s for free

so it's time to update the suggested mission of Jul 07, 2013
including as well the improved script for Earth escape of Jul 27, 2014

  • The future Vega-C puts 3.1t at naturally inclined 400km Leo. Saved 50M$ over Ariane.
  • Sunheat engines raise the apogee to 127Mm radius in less than a year, tilt as needed, end with 2391kg.
  • An O2+H2 engine kicks at perigee to send 1888kg at escaped 4233m/s.
    The optimum is broad, and more speed saves volume under the fairing.
  • 50kg of chemical engine and oxygen tank are dropped. 1838kg remain.
  • Sunheat engines add 5km/s for 9233m/s. End with 1229kg.
  • 300kg of hydrogen tank and sunheat engines are dropped. 1538kg remain.
  • Sunheat engines add 8367m/s for 17600m/s versus Earth. End with 784kg.
  • Jupiter reached in 0.8 years brings the asymptotic speed to 35km/s versus the Sun.
  • The probe reaches 90AU in 13 years after Earth escape.

The remaining 6 D=2.8m engines eject the 754kg in 17 days. They weigh 70kg and the remaining tank 140kg, leaving 574kg for the bus and the science. That's half more than previously with a cheaper launcher - or send 6* more than Vega-C with Ariane.

Several smaller probes could observe the turbulence of the medium, sense it by radio transmissions, explore different directions... Jupiter can spread the probes, including out of the ecliptic plane. The probes can also leave Earth's vicinity at different dates.

The concentrators can serve as antennas. Each can also collect 1W sunlight at 90AU. If relying on propoer orientation, this is more than enough heat and electricity for a probe that makes a measure and transmission per week, but is little for a parallax measurement probe. Spin stabilisation is an interesting option.

Such a mission is the perfect opportunity to test the Pioneer anomaly

Marc Schaefer, aka Enthalpy

Link to comment
Share on other sites

Accelerating nearer to the Sun than Earth is would be more efficient, but if a probe must first brake for that, it gains knowingly nothing. So what if Jupiter sends the probe back to 0.3AU where it accelerates towards the terminal shock?

Alas, the flyby must reduce the eastward speed and lose energy. The leg towards the Sun takes time as well. And the net result is: better than shooting from Earth directly, but not quite as efficient as leaving just by the slingshot at Jupiter.



Having to operate at 0.3AU in addition to 90AU, the probe becomes much harder to design. It also needs a cryocooler for the hydrogen that makes the second kick, while the simple Jupiter slingshot may conditionally live without.

Are there advantages?

The same launch can send an other probe in the opposite direction, after Jupiter puts it on a retrograde orbit. Nice, but it takes about 3 years more, and if waiting 5.4 years, a simple slingshot at Jupiter can target the opposite direction too.

The deflection by the Sun can vary more, by adjusting the perihelion or by pushing earlier or later. My spreadsheet pushes early.

Better: Jupiter can send the probe out of the ecliptic plane, up to polar solar orbits, so the probes can reach the terminal shock in any direction. My spreadsheet computes only prograde and retrograde orbits, but the tilted performance is in between. As the Sun's movement apex is far from the ecliptic plane, reaching this direction may justify the extra engineering and waiting.

Saturn seems to achieve retrograde orbits better than Jupiter does, but the legs to it and back take longer. Flybys at Venus, Venus, Earth before Jupiter may save propellant and achieve better angles. I checked none.

Marc Schaefer, aka Enthalpy

Link to comment
Share on other sites

  • 4 weeks later...

Mass estimates for the 2015 BZ509 mission to bring samples back.

An Atlas V 551, Ariane V, Ariane 64, H-IIB puts 18800kg on a naturally inclined 400km low-Earth orbit. The payload volume demands a bigger fairing, or an own stage, or the whole mass could be reduced.

According to the escape script of Jul 27, 2014

  • Eight D=4.57m sunheat engines bring 14503kg to 127Mm apogee with adequate tilt in about 17 months. Or add engines, 30 kg each.
  • A small O2+H2 engine gives 4233m/s over Earth's gravity and speed to 11450kg.
    If accounting tank masses, the mass optimum would be over 4233m/s, and exceeding this new optimum would save volume.
  • 890kg of O2 and shared H2 tanks and 301kg engine are dropped, leaving 10259kg.
    A 25kN engine with fuel cells and electric pumps would outperform the RL-10.
  • The eight sunheat engines bring 5042kg to 13058m/s over Earth's gravity and speed in 33 days.
  • The tank for 5.2t H2 is dropped, leaving 4108kg heading to Saturn in 2.3 years.

The optional own stage can stop here or already after escape.

I take 182kg per ton of H2 for insulated balloon tanks in trusses that carry a heavy load during launch. Dropping the trusses earlier, for instance with the O2+H2 engine, is uneasy but would save mass.

Here I neglect the 2.5° inclination of Saturn's orbit, but it could cost up to 1.8km/s.

The Saturn flyby is for free.

Hohmann transit to BZ509 at 3.18AU perihelion takes lengthy 8.6 years. Tilting from 180° to 163° takes approximately 3856m/s. Eight sunheat engines optimized for about 6.7AU need 1.2 year to eject 1345kg. The chambers may differ from the ones used at 1AU, which adds few unaccounted kg, but then 100kg of 1AU chambers would have been dropped before.

Four 30kg engines and 245kg tank are dropped, leaving 3332kg heading to BZ509.

Four engines brake by 1331m/s in 190 days because dice limit the sunlight to the power available at 6.7AU. A shorter trip would need more fuel mainly here. 2993kg arrive at 2015 BZ509 full of nice tools and toys.

I take Saturn at aphelion, BZ509 at perihelion when arriving and at aphelion when leaving. Lengthy orbits don't let choose, and I didn't check the consequences of BZ509 being far from its nodal points.


Of the 2993kg, 300kg are a return capsule, 100kg a bus for the return leg, 100kg are four sunheat engines kept for the return, 110kg the tank with 236kg H2 for the return and 339kg already used upon arrival.

The other 2147kg comprise 400kg of bus and 1747kg to split among remote sensing and landers:). Some robotics catch the landers and transfer the samples.

As I suggested elsewhere, remote sensing could include a pulsed laser powered by the sunlight concentrators, hydrogen and xenon jets to erode the surface, a hydrogen gun for deeper sensing, maybe tethered hollow harpoons to take samples without landing. I'd prefer several landers of different construction for redundancy: landing, anchoring and sampling hardware...

The size, shape, mass and composition of BZ509 are unknown. Reflection suggests D=2km, then iron-nickel would weigh 3*1013kg. Take-off would need 2.2m/s provided by springs, possibly hydraulic, and the ferry might orbit below R=3.5km but its sunheat engines couldn't levitate it. But if BZ509's mean density is 500kg/m3, the Lagrange point is at R=0.9km, so no orbit is possible. As with Chury, staying near BZ509 is difficult, until someone has an idea.

Between the nodal points, the mission has 5.8 years to take probes, optionally more for remote sensing.


858kg leave BZ509's vicinity. The four sunheat engines use 236kg in 148 days at 7.09AU to brake by 4000m/s within the retrograde orbital plane to join Earth directly.

If I misunderstood and detilting by 17° is needed, the sunheat engines still achieve it, but the capsule must be half as heavy and the leaving aggregate twice as heavy. Not the hypothesis here.

Not forgetting the fuel for fine-tune would be better. The bus, engines and tank separate from the 300kg capsule that re-enters Earth's atmosphere at 68km/s.

A capsule has to decelerate by ~500g or it would exit the atmosphere, ouch. Some flat form for L/D=1, if feasible for that speed, would reduce it to 74g downwards and backwards, or total 105g.

The 300kg capsule comprises:

Marc Schaefer, aka Enthalpy

Link to comment
Share on other sites

  • 1 year later...

This is how I imagine the development of the sunheat engine.

----- Chamber -----

It needs a very small team, with at least one scientist of very broad knowledge, and 1-2 more people used to innovate. A team of more specialized people may achieve it, or not, and less well. A bigger team would not work proportionally faster.

A experiment rig is indispensable, available exclusively to the team all the time. Reformulated: computer simulations followed by a short experimental verification at an external facility would fail.

The concentrator used to develop the chamber provides the power, power density, convergence angle of the flight model. It differs by the mass, materials, fabrication, lack of deployment.

----- Concentrator -----

It needs at least one mechanical designer used to space hardware and innovative designs. Expertise for optics, heat and control could come from a separate brain if needed. Expertise for launch environment, resistance of structures, space environment, deformations in operation, materials, fabrication processes should be available in the designer's brain. A team of more than 2 people would not work proportionally faster.

Both teams can share people. The concentrator development can start immediately or once the chamber shows convincing results.

----- Tanks -----

The vacuum-insulated tank technology I proposed doesn't still exist, though Ariane 6 probably adopted it and future aeroplanes too. It's a significant development too that needs a specific test ground and can be sold separately as hardware or technology. Very few people, no banal knowledge, one year.

Missions to the outer space need for the hydrogen a cryocooler I described elsewhere. It should have existed for decades at least for oxygen, but was neglected. The uneasy development is less exotic than for the chamber. Can be sold separately as hardware or technology. Few people, 2-3 years.

Missions to Geo, the Moon, conditionally Mars, need no cryocooler, nor does a demonstrator of the sunheat engine in Leo.

----- Company -----

If a company is created especially, the burden needs at least half a seasoned person for months.

Selling the engine or the technology needs later a salesperson who knows the people and habits of the space activity.

The lab for a Vega-sized demonstrator can be normally tall with a good area and wide doors. For Ariane or Falcon 9, the lab must be 3 stories tall. Hydrogen, strong lamps and 24h tests need detectors and access control.

----- Dough and proof in space -----

All these developments are rather affordable. I evaluated in 2018 in a rich country to 400k€ the development of the chamber alone.

Scaling for varied launcher sizes needs limited development.

Add the concentrator and the tank, and for 2-3M€ you define the future of space travel.

- BUT -

Customers want flight-proven hardware before they design a mission around it. The sunheat engine can't be very small, and mission planners may want the real size proven in flight. Sharing a launch isn't easy neither, as the definitive concentrators are as wide as the fairing enables.

Problem: a launch on Electron costs 6Musd, on Vega 20M€, well over the development cost.

This is a toxic combination for the startup company I had imagined. Unless a small cheap seat can be found on a launcher (which is technically difficult), the business model of a small company selling hardware looks now bad to my eyes.

Maybe an agency or bigger company shall develop these technologies, or a smaller company sell them the technologies before they have flown.

Marc Schaefer, aka Enthalpy

Link to comment
Share on other sites

Thoughts about the chamber's experiment rig here under.

I put far less time in the concentrators and don't plan to detail their development. The cryocooler is small, it needs experimentation in a normal lab. Thoughts may come or not, rather elsewhere, about the test ground for the multilayer-insulated hydrogen tank.

----- Light source -----

It must reproduce concentrated sunlight in space: same power per surface unit, at the same convergence angle that distributes the heat in the chamber. Sunlight on Earth can't. A xenon arc lamp is the usual and possibly only solution. D=2.8m concentrators for Vega-C need 8kW light for one Sun-Earth distance, D=4.57m for Ariane and equivalents need 22kW light. Single lamps achieve this.

Extreme temperature for months defines much the chamber design, so the rig must test it. 24h operation is needed despite hydrogen and strong light. Spare lamps at hand are prudent. Cheap electricity is better.

To protect humans and limit heat at the test vessel's porthole, I prefer that the lamp or a filter suppresses the UV. With identical power, the difference at the chamber is negligible.

----- Concentrators -----

An elliptic mirror shall concentrate the light from the lamp to the chamber being developed. It could be smaller than the flight concentrator but must distribute the power faithfully over the directions. Successive mirrors may help.

Test of the real concentrator needs an additional big parabolic mirror that makes the lamp's light parallel. The vessel can then sit in between if the focal distance is meaningful without secondary and tertiary mirrors.

The chamber can be tested with the optical path fixed, and the optical path steered without the chamber. Acceptable to me. Or in the optics of Jun 21, 2013 here
the axis from secondary to tertiary mirrors can be vertical, the primary rolls around its axis and the vessel and chamber rotate around the secondary to tertiary axis, with all piping and wiring.


----- Vacuum vessel -----

It hosts the chamber, maybe a last mirror, but not the primary, to save costs. Light enters almost concentrated through a porthole. More ports are desired for observation, assembly, and varied optical setups. The vessel must host the nozzle, D=0.37m for D=4.57m concentrators.

The windows should be coated against reflections, absorb little the remaining wavelengths, and can receive blown air on the outer face. A flat window may alter only the focal distance but with chromatic aberration, while a spherical one centred on the focus could be slightly astigmatic, but I'm weak on optics. A diffusing diaphragm around the window is safer.

Water cooling seems necessary in the vessel, to protect the walls in a faulty operation, and to cool the ejected hydrogen, whose stagnation temperature would fit tungsten only.

A chamber for D=4.57m concentrators ejects 0.2g/s hydrogen and 0.08g/s for D=2.8m but pressure shall remain <<1Pa. That's a difficult task achieved by a big turbomolecular pump, not cheap. Second hand? An actively cooled convergent downstream the nozzle would help much. A primary pump is required too.

----- Hydrogen -----

0.2g/s over a week-end is 43kg hydrogen, 500m3 at STP, or many 300bar bottles, dangerous. Over 3 months it's 1.5t, expensive and inconvenient. I couldn't find pleasant ways to produce slowly in situ neither. Better reuse the hydrogen pumped from the vessel. This may need some purification. The total hydrogen amount is then small.

When switching between air and hydrogen, I'd purge all parts with nitrogen, which needs additional circuits and bottles. Ask an expert.

----- Fabrication -----

The mirrors and the lamp mount could be subcontracted, because people protection, surveillance sensors, and distribution over spectrum and angle aren't trivial.

The vessel could be subcontracted. Or adapt a second hand vessel to the task.

Hydrogen piping, with the purge design, sensors and certifications, could be subcontracted.

Marc Schaefer, aka Enthalpy

Link to comment
Share on other sites

  • 3 months later...

Nasa and Esa consider a mission to bring Mars samples back to Earth. The sunheat engine makes it much easier and enables a more ambitious mission.

The artist rendering I saw recently suggest ion propulsion with big solar panels and small thrust. The sunheat engine brings a specific impulse similar to ion engines and a much stronger thrust for the same sunlight collecting area, so a heavier spacecraft can manoeuvre faster.

I won't detail that.

Link to comment
Share on other sites

  • 1 month later...

More people think at the sunheat engine: the Johns Hopkins University's Applied Physics Laboratory designs a craft and mission to reach the heliopause in 25 years, nice
popularmechanics.com - wired.com - jhuapl.edu

I just hope that their readers won't remember APL's sunheat engine design as the only possibility.

The APL's design works at 3 solar radii. A concentrator lets my common design work up to Saturn.

Less good materials limit the temperature at APL's engine. They claim 600s theoretical specific impulse, while only-tungsten plus low hydrogen pressure give my design 1200s.

Stronger Oberth effect nearer to the Sun is an advantage for APL's mission, but my engine and mission reach the heliopause in 10-15 years
scienceforums - scienceforums - scienceforums

Edited by Enthalpy
Link to comment
Share on other sites

  • 7 months later...

Nasa's Europa Clipper mission to Jupiter's moon Europa progresses
that's fantastic news for all people interested in science. The possible liquid ocean below Europa's ice shelf is a potentially liveable location in our solar system.

As decided recently, a Falcon 9 heavy will start the craft that, after a 6 years trip with gravity assits, will orbit Jupiter with 2+t and fly quickly past Europa many times. Chemical propulsion (and radiations, they say) needs the launch performance and limits the final orbit. Flybys prevent to land, make long-base seismic images, study finely the magnetic field, make detailed maps.

As opposed, my sunheat engine needs only a standard Falcon 9 launch and puts quickly 3.4t in Europa orbit
there, the sunlight concentrators provide 6kW electricity instead of 300W for Europa Clipper and double as high-gain antennas, answering the transmission argument.

Landers could bury their electronics under local materials, dust, ice, as a radiation shield if any necessary.

We should have had the sunheat engine for decades. But it's still time to develop it.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.