I am building a world where one side of the planet has normal progression of day and night whereas in the other hemisphere it is always night.

At first I had no idea how this could be possible in the real world. But then I came up with this idea:

The world of the story is in fact a moon, orbiting a gas giant. With the particularity that one hemisphere of the moon it's always faced to the system star. Because of this, the night is caused due the gas giant shadow. And the back hemisphere is always at night (with some light reflected from the gas giant).

I have hand drawn a schema, it's not a Picasso but I hope you understand me (the drawing is not on scale).

enter image description here

It wasn't until I made the drawing that I realized that the night in the frontal hemisphere would be too short. The rotation time around the gas giant in a 24 hours period could be a problem too.

Can this system exist? Can this system exist whit day/night periods similar to Earth? If we want prolongate the night, how could result the system?

I have no problem adding other elements like rings, other moons and so. I let this to your imagination/knowledge. But less is more. Thanks for reading, waiting your answers.

  • 1
    $\begingroup$ Just a comment: the systems with tidally locked planets having a always-day and a always-night hemisphere tend to have rather unpleasant weather, as one hemisphere is significantly colder than the other. And by "rather unpleasant" I mean "way to harsh to sustain life". Your world would have similar problem (although to a lesser extent) $\endgroup$
    – Mori
    Commented Aug 22, 2019 at 12:39
  • $\begingroup$ I hoped that the weather could be softened with the excuse of ocean currents. $\endgroup$ Commented Aug 22, 2019 at 12:48
  • $\begingroup$ Are you married to the idea of it being a moon and not a planet? $\endgroup$
    – John
    Commented Aug 22, 2019 at 14:48
  • $\begingroup$ @Mori that only applies to a planet tidally locked to its host star. A moon tidally locked to a planet would not be subject to such a problem. $\endgroup$
    – stix
    Commented Aug 22, 2019 at 17:57
  • $\begingroup$ @John No, there is not need to be a moon. It is just because it's the only idea i have. Any answer is welcome. $\endgroup$ Commented Aug 22, 2019 at 22:14

7 Answers 7


Geometrically speaking, you need the moon to have its rotation synchonized with the planet's revolution around the star (thanks to David Robie for the vocabulary clearup). I believe this is possible, but highly improbable, in a story I would probably perceive it as too improbable and deux-et-machineous, unless there is some hand waving explanation involved (something happening in the past locking the moon rotation on such improbable speed). Note that the moon revolution around the planet would have a slight effect on day/night line and on the position of the sun in the sky (though I am not sure if strong enough to be observable).

Also, there is the question of the harsh weather. I would guess that the "day-side" being partially shadowed by the gas giant the temperature gradients would be significantly more manageable that in the case of pure tidally locked planet, but it still does not really sound like a nice world to live on.

TlDr: I am afraid that without some magic/overlooking aliens with god-like technology/something there is not much plausibility in the story.

Edit: one more thing to consider: the gas giants tend to be too far from the star for the system to work as described.

  • 2
    $\begingroup$ It has been calculated that the year of a planet has to be at least 9 times the length of the orbit of the moon around the planet for the moon's orbit to be stable. Thus the Moon's orbit around the planet and it's tidally locked rotation rate can not be synchronized with the planet's orbit around the star. $\endgroup$ Commented Aug 22, 2019 at 16:24

A gas giant satellite will most likely be tidally locked (which kinda defeats your purpose) or will at best have, for just a few millions of years, a rotation period equal to its orbital period.

I think similar (but not duplicate) questions have been asked before, and the conclusion is always that no orbital arrangement would allow for it.

Let me propose a different idea: if one hemisphere has a vortex over it and a source of fine ash (say, an active volcano), the ash trapped in the atmosphere would block sunlight. This might not cover a hemisphere with surgical precision, but may get you close to what you want.

  • $\begingroup$ "few millions of years" sounds enough to me. But what about the periods of night/day? The solution you gave to me could be a good one. But the problem is that stars can not be seen and that's a big problem for the story of the inhabitants of this world. $\endgroup$ Commented Aug 22, 2019 at 12:12
  • $\begingroup$ @MiguelNoTeimporta for a gas giant like Jupiter, your days (full rotation) would be from 9 to 12 hours long. $\endgroup$ Commented Aug 22, 2019 at 12:16
  • 1
    $\begingroup$ @Renan how so? Io (Jupiters moon) has orbital period of 1.75 days (other big moon even longer) $\endgroup$
    – Mori
    Commented Aug 22, 2019 at 12:52
  • $\begingroup$ @MiguelNoTeimporta please see Mori's comment above. Mori is right - I had some moons from Saturn in mind when I gave out those numbers above. Your arrangement may be somewhat possible with a 24h day. $\endgroup$ Commented Aug 22, 2019 at 13:02

Your planet is in a close orbit around a rapidly rotating supermassive black hole. Not so close that the CMB is blue shifted into a source of permanent daylight, but close enough that gravitational lensing causes the image of the event horizon to fill most of the sky in one hemisphere. That's your permanent night.

A small companion star provides a day/night cycle to the other side facing away from the black hole.

  • $\begingroup$ This is a pretty cool and badass alternative. $\endgroup$ Commented Aug 22, 2019 at 22:22

It is possible to set up such a situation, at least temporarily. Probably not long enough for life to evolve. But maybe it hopped across somehow from somewhere more stable. In the Jupiter system there are a lot of big objects a lot closer. Maybe it's easier.

You will need to be thinking about relative size of the gas-giant planet and the moon's orbit around it. You need to do some actual orbit calculations to get some idea of how long the eclipse lasts. As a lowest-order estimate, you could take the relative size of the gas-giant planet and the moon's orbit. The part directly behind the gas-giant is the only part in shadow.

It's more complicated than that as well. The moon is unlikely to have a perfectly circular orbit around the gas-giant. And the gas-giant is unlikely to have a perfectly circular orbit around the star. This will mean that the dark portion of the orbit is different from orbit-to-orbit in a complicated pattern. It would be some non-trivial (though possible) amount of work to predict that pattern. There's probably an app-for-that somewhere on the net.

Like so: When the in-shadow part of the orbit corresponds to the part that the moon is farthest from the gas-giant, it is moving slowest, but has the smallest apparent angle for the gas-giant. When it's the closest part, it's moving faster with the largest apparent angle. You would need to work out the net effects of this, using reasonable values for the mass and orbits of the various objects. Don't forget the Roche limit. It might be reasonable to start with the Jupiter system and see what you can put together. This website gives you the basic data of orbit period and radius.

Also, the moon's orbit and orbit of the gas-giant are unlikely to be perfectly in the same plane. If the moon rises above or below even a little, it could easily miss the gas-giant's shadow most of the time. Maybe you miss night for 3/4 of the gas-giant's orbit period.

So the moon Io has an orbit of 1.769 Earth days, and orbit radius of 422,000 km. Comparing that to Jupiter's radius of 69,911 km, the naive calculation means Io is in shadow no more than about .053 of the orbit, or about 2 and a quarter hours. (Geometry left as homework.) Moons further out would have longer periods, but smaller fraction of their orbit in shadow. So Io is probably the closest you can get. And Io is tidally locked to Jupiter, so it's not your planet.



This is really difficult, but I think there might be one or two reasonably plausible natural situations, and at least one artificial situation created by an advanced civilization, where one part of a world has alternations of light and dark and another part has only eternal darkness.


To appreciate some of the problems with such requests, you might want to look at an attempt to design a vaguely similar situation in these two posts:





All planets and other astronomical bodies rotate at various speeds determined by various factors.

So if a planet orbits around a star close enough to get enough light and heat from that star to be warm enough for life, presumably like Earth life, it will have days and nights. At any one moment half the surface of the planet will be receiving light from the star and the other half will be in shadow and dark. As the planet rotates every part of the planet's surface will alternately experience day and night.

If the axis of the planet is tilted the relative length of night and day will vary with the seasons at different latitudes. But even at the poles periods of constant light or constant night must last less than half of the planet's year. So no spot on that planet can have either eternal light or eternal night.

Except that the distance from the planet to the star where the planet can have the right temperature for life varies with the luminosity of the star, and the luminosities of main sequence stars vary according to their mass. And a small change in the mass of the star will cause a much larger change in the luminosity of the star.

So a small mass star will be really dim compared to the Sun, and thus the habitable zone around that star will be really close to it. Any planets orbiting the habitable zone around a low mass, dim star will orbit really close to it. And thus the tidal forces on the planet from that star will be very strong and the planet will become tidally locked to the star very soon in geological time.

As the planet becomes tidally locked to its star, its rotation rate will be slowed and slowed until the time it takes for the planet to make one rotation will be the same as the time it takes for the planet to orbit around the star one time. Thus one side of the planet will always face the star and will have eternal day, and the other side will always face away from the star and will have eternal night.

So how do you get the side facing the star to have alternating day and night, while the other side has only eternal night?

That is very tricky.


The original question asks for a moon orbiting a planet, with the moon being tidally locked to the star that the planet and moon both orbit. But calculations indicate a moon would have to be tidally locked to its planet instead of to its star. it is considered impossible for a moon to be tidally locked to its star instead of to its planet.

One could try making the planet a planet sized moon orbiting a giant planet and tidally locked to the planet, and not to the star. Then the side of the moon that always faced away from the planet would have alternate day and night. The side that always face toward the planet should also have alternate day and night, except that its day should be interrupted by a long eclipse every day when the moon passes into the shadow of the giant planet. And if the giant planet has a bunch of other large moons, the moon in question might often be in eclipses caused by them.

So I suppose that it is theoretically possible that someone could design a plausible configuration where at at least a small segment of the side of the moon that always faces the planet would always be in eternal darkness.


Another possibility is a planet orbiting in the habitable zone close to a dim star, and tidally locked to that star, as discussed above. Except in this case the star is a close binary, a very close binary, and the planet orbits around both of the stars.

Astronomers have theorized about planets in such orbits, called P-type or circumbinary orbits, and have detected some. It is possible that two really dim stars could orbit close enough to each other that a planet in their combined habitable zone would orbit close enough to be tidally locked to the stars. Thus one side of the planet would have eternal night. So what about day and night on the side facing the two stars?

The planet's orbit would probably be dragged by tidal forces to orbit the two stars in the same plane as they orbited around each other. So each of the two stars would periodically eclipse the other one as seen from the planet - they would be an eclipsing binary as seen from the planet.


And if one of the stars is a vary dim star, or less than a star, it would contribute very little light to the planet, and when it eclipsed the bright star a sort of night would fall on the star facing side of the planet. So the dimmer object should be cool enough to emit very little visible light, and also large enough in diameter to totally block off the light of the star when eclipsing it.

So the dimmer object should be a "puffy planet" or a brown dwarf.



And I guess that various possible configurations of such a system might have eclipses on the star facing side of the planet last for about 0.1 percent of the time, or 1 percent of the time, or 10 percent of the time, etc., etc. That would not be like having equal periods of day and night, but it would be a lot different from having eternal day on that side of the planet.


A rogue planet is a planet in interstellar space instead of in a star system.


We can expect rogue planets far from any star to be very, very, very cold. But an advanced civilization could terraform a Earth-sized rogue planet and make it habitable. There would be many possible methods to heat it to the desired temperature.

One such method would be to create an imitation of the geocentric model of the solar system, with a "sun" orbiting the planet instead of the planet orbiting the "sun". The advanced civilization would build a gigantic artificial sun" satellite to orbit the planet, with countless gigantic fusion generators to power countess gigantic lamps aimed at the planet.

The artificial "sun" satellite would orbit the rogue planet, creating a repeating "sunrise", day, "sunset", and night sequence over every part of the planet.

But if for some reason the highly advanced civilization wants to create a world with eternal night on one side, and eternal day on the other side, they can put the artificial "sun" satellite in a synchronous orbit so it is always above one place on the planet. So one side of the planet will have eternal day and the other side of the planet will have eternal night.

But if for some reason the highly advanced civilization wants to create a world with eternal night on one side, and alternating day and night on the other side, they can put the artificial "sun" satellite in a synchronous orbit so it is always above one place on the planet, and then periodically turn the artificial "sun" satellite on and off. So one side of the planet will have alternating day and night and the other side of the planet will have eternal night, as requested.


As in most physics, the answer is absolutely.

Consider Kepler's laws. One of which states that any sectors which taken at a point perpendicular to the plane of orbit, and have the same area, will have the same time to traverse for a single body orbit.

Consider the moon, which is tidally locked to the red dwarf. The Planetary body casts a shadow like so:

enter image description here

So long as A1 and A2 are the same area, the day-night cycle will remain the same times from each other. The only parameter left is the orbital period. In this case you mentioned that It's earth-like, so the period would be 24 hours.

This example, however, is rather unrealistic, but one can ground the premise more in reality by simply making the moon's orbit more eccentric: enter image description here

The law stays the same, and so long as the areas divided by the rays A1, and A2 are equal, then the day night cycle will hold.

Note that the planetary body in this case is comically large, and that to have such a cycle in the Goldilocks zone, that the atmosphere would need to be around 0-125% thicker to trap heat. Alternatively, you can make the orbit much smaller, but keep the proportions (eccentricity) of the orbit the same, if this is done, then the mass of the blue planetary body must increase.

  • $\begingroup$ I wouldn't thought about elliptic moon orbits but as you explain looks viable to me. $\endgroup$ Commented Aug 22, 2019 at 22:24
  • 1
    $\begingroup$ The ellipse will not remain aligned with the sun throughout the year. $\endgroup$ Commented Aug 22, 2019 at 23:36
  • $\begingroup$ @LoganR.Kearsley it will if the parent planets orbit is slow enough. $\endgroup$
    – tuskiomi
    Commented Aug 23, 2019 at 1:56
  • 1
    $\begingroup$ @tuskiomi It's not a matter of being slow enough, it's a matter of having all of the torques on the moon's orbit from all other bodies in the system being balanced exactly right. The planet has to orbit at exactly the right speed, the moon has to orbit at exactly the right speed and exactly the right distance, the star has to be exactly massive enough... keeping the moon's orbital nodes aligned with the star at all times is just a huge house of cards. $\endgroup$ Commented Aug 23, 2019 at 2:16
  • 1
    $\begingroup$ @LoganR.Kearsley I'm not sure you understand the question. The question is: Is such a scenario possible? The answer: Absolutely. The question is not How possible. The question is not How likely. The question is: Is such a scenario possible? $\endgroup$
    – tuskiomi
    Commented Aug 23, 2019 at 14:05

It's possible. You could tell people that the moon's rotational velocity matches its revolution veolicity around the planet. That is to say, the dark hemisphere is always facing the planet away from the star and never receives light.

Edit: Mori and I talked it through. They are correct in saying that you can also accomplish this by synchronizing the moon's rotational period with the planet's revolutionary period.

  • 1
    $\begingroup$ The problem is that the moon's rotational velocity does NOT match its revolution velocity around the planet, but rather the velocity of planet around the sun. Unlike the first the latter is highly improbable (though I dare to say possible) arrangement. $\endgroup$
    – Mori
    Commented Aug 22, 2019 at 12:33
  • $\begingroup$ @Mori What I'm thinking is that the moon has one side which is always oriented in the absolute direction of the star, while the other faces away from it. The day night cycle on the moon is caused by it travelling through the planet's shadow. Your description of the rotation matching planetary revolution about the star just means a moon that spins really fast. Did you mean having the moon-planet revolution match the planet-star revolution? Wouldn't that just permanently put the moon in the planet's shadow, meaning neither hemisphere receives light? $\endgroup$ Commented Aug 22, 2019 at 12:48
  • $\begingroup$ Yes, I am thinking the same thing. I believe to achieve this the moon must perform one revolution around its axis per planetary (gasgiantic) year (might be wrong though, now you have got me seriously confused) $\endgroup$
    – Mori
    Commented Aug 22, 2019 at 12:57
  • $\begingroup$ Things rotate around their axis and revolve around other objects. If it makes one rotation per planetary year, then you have light exposure to all sides at some point during the year as it revolves around the planet. My answer is that as it revolves (more surface area leaves the shadow of the planet), it rotates (turning more of the light hemisphere to face the star) so that no part of the dark hemisphere ever faces it. $\endgroup$ Commented Aug 22, 2019 at 13:07
  • $\begingroup$ As for vocabulary: my bad, sorry. As for the geometry: I still do believe that there needs to be one revo... I mean rotation of the moon per planetary year. Revolution of the moon around the planet is completely other thing. $\endgroup$
    – Mori
    Commented Aug 22, 2019 at 13:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .