In 2019 Aug 27's test flight, during the final 8 seconds before landing (T+00:45), Starhopper's plume was suddenly much brighter. Why?

The Raptor's thrust can't have changed much then, because Starhopper's speed and trajectory hardly changed. A different fuel mix? (Again, why bother?)

The glow starts about when the visible plume touches the landing pad, which had already started to glow. But how could that touch affect the appearance of the plume itself? Burning concrete dust (or whatever) can't propagate up towards the engine.

  • 2
    $\begingroup$ Burning metal from inside the engine? $\endgroup$ Commented Aug 27, 2019 at 22:55
  • 8
    $\begingroup$ Interesting question! Just fyi SpaceX edits their YouTube videos after some period of time to remove segments with no activity, so it's always good to include a time from the launch clock as well. I've edited the time in your clip so that the video plays properly now that they've dramatically shortened it. $\endgroup$
    – uhoh
    Commented Aug 27, 2019 at 23:30
  • $\begingroup$ To my eyes the speed changes drastically about a second after the brightness increases. $\endgroup$
    – user32989
    Commented Aug 29, 2019 at 14:07
  • $\begingroup$ Scott Manley's video at 3:25 show that the glow starts half a vehicle length below the nozzle, then progresses to the nozzle within a second. That supports dust entrainment, and casts doubt on engine crud burning. At 7:58, the excerpt from Tim Dodd's camera shows a few yellow puffs, but no sustained glow before the vehicle is deep into the dust. So I still claim dust entrainment. $\endgroup$ Commented Sep 2, 2019 at 0:55
  • 1
    $\begingroup$ The engine may have failed, but if that were the reason for the yellow color, than one should also assume that it has failed also on takeoff and that for some unknown reason it was not making the plume yellow during the horizontal flight. Dust is a simpler explanation. $\endgroup$
    – BlueCoder
    Commented Sep 2, 2019 at 8:35

4 Answers 4


Indeed it is due to contaminants — dust! — in the plume.

Did you notice all the dust being kicked up around the landing site? Some of that dust cloud, a very small and low-dust-density part of it (so it's really hard to see), flows upward, then back toward the rocket, and then is entrained in the rocket's exhaust plume. It's just like the toroidal flow around a helicopter hovering near the ground.

The nice, clean blue plume from the Raptor engine results from efficient burning, not producing a lot of hydrocarbons in the exhaust. The reaction is essentially complete when the exhaust exits the nozzle, so you get the blue of the relaxation emission from excited-state water, and the invisible IR emissions from CO2. When dust is entrained in the plume you get a wide range of chemical species — hydrocarbons, silicates, all the stuff in Texas dirt! — reacting in the plume, and they emit light all over the visible spectrum. Hence the change in color, and conversion of more of the plume's energy to visible light, so the plume brightens.

  • 6
    $\begingroup$ Sounds plausible, but do you have a reference? $\endgroup$ Commented Aug 27, 2019 at 23:32
  • 7
    $\begingroup$ @RussellBorogove, No, this is based on knowledge gleaned from working with folks who did past robotic Mars landings: the behavior of plumes and the airflow around them during landings. $\endgroup$ Commented Aug 27, 2019 at 23:36
  • 5
    $\begingroup$ So that's why the plume isn't bright on launch also. Then, the vehicle's ascent needs more thrust, so the toroidal vortex is spreading out too fast to reconnect with the plume. $\endgroup$ Commented Aug 28, 2019 at 1:32
  • 10
    $\begingroup$ Scott Manly just posted a video where he was presented with some arguments against this explanation, namely a quick flash beforehand and the fact that it's very bright all the way back to the nozzle. $\endgroup$
    – JVal90
    Commented Aug 28, 2019 at 17:07
  • 1
    $\begingroup$ When watching the video, I immediately thought it oculd be due to deeper throttling of the engine prior to landing. When taking off, the flame was the standard Raptor-blue/pink we're used to seeing, so I doubt it was dust. Perhaps they ran the engine fuel-rich for that portion of the flight? I know Raptor can throttle down with optimal fuel/oxidizer ratio, but maybe they wanted to pre-cool the engine or maybe even form some soot on it to prevent engine parts from being oxidized after landing? $\endgroup$
    – JohnEye
    Commented Aug 29, 2019 at 12:44

The bright yellow light from RP-1 burning engines is from carbon soot built up from already-complex carbon molecules in kerosene. However the single-carbon methane molecules quickly oxidize to CO2 and H2O and are not conducive to soot growth. When we cook using natural gas (methane) we rarely see soot produced, and so the flame is blue mostly due to diatomic molecular carbon C2 (see the vibronic spectrum in this answer and this answer) and possibly excited water as @TomSpilker mentions and as discussed in What is the cause of the blue light from LH2/LOX rocket engines?.

I like his explanation that the bright yellow light is from dust being entrained in the plume, but I don't think the light is chemical in nature.

Instead, it's probably blackbody thermal radiation from the dust (bits of rock) either solid or melted, that's instantly heated to the exhaust's temperature. The exhaust gas can not efficiently radiate because it's not a blackbody, but the moment the particles enter the plume they can readily glow.

That's why its so bright and so uniform in the same rocket-exhaust yellow color we see from soot-producing LOX/RP-1 plumes.

  • 1
    $\begingroup$ I agree, it's not all from chemical reactions, a good fraction of it is indeed blackbody radiation. It's a mix: there's a fair portion of biologically-produced species in the Texas soil, and at those temperatures they react like crazy, yielding some species with discrete emission spectra, and some (like soot) that also radiate via blackbody radiation. $\endgroup$ Commented Aug 27, 2019 at 23:46
  • 3
    $\begingroup$ @RussellBorogove oh no, I've accidentally revealed my newly-invented LOX/LCO engine to the world! $\endgroup$
    – uhoh
    Commented Aug 27, 2019 at 23:55
  • 3
    $\begingroup$ If only one could accept two answers! Yes, soot blackbody radiation explains the color, but Tom explained why the plume gets brighter at all. $\endgroup$ Commented Aug 28, 2019 at 1:39
  • 4
    $\begingroup$ @CamilleGoudeseune being right next to the beach I'd bet a lot of money on sodium in the sea water being the reason for the orange color. The sand will be full of it and even small amounts are enough for a bright orange flame color. $\endgroup$
    – Christoph
    Commented Aug 28, 2019 at 11:06
  • 1
    $\begingroup$ "The exhaust gas can not efficiently radiate because it's not a blackbody" -- can you explain why? $\endgroup$
    – Yakk
    Commented Aug 28, 2019 at 14:07

Scott Manley comments in his video that it could be part of the engine has started to burn or erode, introducing new elements otherwise very pure rocket exhaust.

Evidence for this (as opposed to entrained dust) is that the yellow colour starts suddenly, late in the flight, and it starts right up in the engine. If it was dust, you'd expect it to start at the bottom of the flame and move upwards, and to occur on take-off (as the vertical speeds were similar). Also the other flame visible at the top of the bell (just before the plume goes yellow), the hard landing and the tank that broke loose.

I guess we'll have to wait for an official SpaceX answer to know for sure.

  • $\begingroup$ This is my suspicion as well, as I commented earlier. $\endgroup$ Commented Aug 29, 2019 at 11:44
  • 3
    $\begingroup$ If I have followed the accepted answer correctly, we would expect the dust to start at the top of the flame. The (hot) rocket plume goes down, hits the ground, spreads out, kicks up dust, and then starts to rise (because it is hotter than the surrounding air). This causes the dust to rise into the air (but not into the plume - that is travelling down at near supersonic speeds). The dust spreads out, and some of it will spread towards the rocket. At this point it gets entrained - and we have the glow. $\endgroup$ Commented Aug 29, 2019 at 11:57
  • $\begingroup$ The formal studies of dust with rocket plumes that I found were only for the lunar surface. No vortex, no entrainment. But in atmosphere, the dust should get sucked into the low pressure region next to the nozzle, not into the high pressure region at the ground. $\endgroup$ Commented Aug 29, 2019 at 14:56
  • $\begingroup$ If dust was being sucked up the exhaust, wouldn't we see it in the video? $\endgroup$ Commented Aug 29, 2019 at 15:09
  • 1
    $\begingroup$ This is the most likely answer. I would add that the landing was harder than expected, given that the feet all broke on impact. Perhaps this was due to engine degradation. Or perhaps the flight computer landed earlier than expected with greater weight of unused fuel... perhaps a combination of issues $\endgroup$
    – Coomie
    Commented Aug 30, 2019 at 6:57

The exhaust is clean when leaving the engine, only turning yellow when it mixes with dust; originally posted by Elon Musk on Twitter. It is just dust. If you look at the exhaust when it leaves the engine, it is clear. It only turns yellow further down, when it mixes with the dust. I don’t know what the small flame is on the bottom of the Starhopper though.

  • $\begingroup$ Which T+xx:xx is this image from? How does this answer expand on the others? $\endgroup$ Commented Sep 2, 2019 at 2:03
  • 1
    $\begingroup$ Elon Musk posted it on Twitter. It was speculated here and elsewhere that the changed exhaust color originated where the exhaust leaves the engine, and that the engine might have started consuming itself. But in this photo you can clearly see that the color change is not where the exhaust leaves the engine, but lower down. Evidently the color change is due to combustion of the dust kicked up when it is hit by the extremely hot exhaust gases. $\endgroup$
    – Rachel
    Commented Sep 3, 2019 at 5:15
  • 2
    $\begingroup$ After a second start-and-stop viewing of Tim Dodd's slo-mo video it appears there are multiple things going on. For most of the last few seconds of the flight indeed the yellow-orange tinge to the plume is as seen in the image above. But at other times, such as at 4:52, 5:00, and 5:07 in Tim's video, there appear bright downward-propagating streaks that start at the nozzle skirt. Then at 5:38 there's a very bright white streak/flash, again starting at the skirt, followed 2-3 seconds later by the orange flame from above the nozzle. I don't think these are all results of the same thing. $\endgroup$ Commented Sep 3, 2019 at 7:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.