Orbital Mechanics
of Space Mirrors
Why can't a space mirror hover over a city? Why doesn't geostationary orbit work? How does sun-synchronous geometry enable the terminator window? The physics constraining space mirrors — worked through from first principles.
Why Can't a Space Mirror Stay Over One City?
This is the most common question about space mirrors, typically prompted by coverage of the 2018 Chengdu "artificial moon" proposal or by the claim that a mirror constellation could provide "continuous" city illumination. The answer is straightforward but worth working through precisely.
A satellite in orbit is not hovering — it is in free fall, perpetually falling toward Earth while moving sideways fast enough that it keeps missing. At any given altitude, there is exactly one orbital speed that maintains a stable circular orbit. Slow down, and you fall into a lower orbit (or re-enter). Speed up, and you climb into a higher orbit. There is no LEO orbit in which a satellite can maintain a fixed position over a ground point while Earth rotates beneath it.
At 500 km altitude — a typical LEO for space mirror proposals — the required orbital velocity is approximately 7,612 m/s (7.6 km/s). The satellite completes one orbit in about 94.5 minutes, crossing 27,500 km of ground track per orbit. From any fixed point on the ground, the satellite rises above the horizon, crosses the sky in 3–5 minutes, and sets again. The next overhead pass occurs roughly 90 minutes later — but the Earth has rotated beneath the satellite's path by then, so the next pass is displaced roughly 1,500 km to the west (at mid-latitudes).
The only orbit where a satellite remains stationary over one ground point is geostationary orbit. And geostationary orbit is physically incompatible with useful space mirror operation — for reasons examined in the next section.
Why Geostationary Orbit Doesn't Work
LEO · USABLE FOR MIRRORS
GEO · NOT VIABLE FOR MIRRORS
Geostationary orbit at 35,786 km would give a space mirror a fixed position over one ground point — solving the pass-duration problem entirely. But the beam geometry makes it useless for illumination.
The minimum size of a reflected sunlight beam is set by the Sun's angular diameter of 0.53°. At any distance d from the mirror, the beam footprint diameter equals d × tan(0.53°) × 2, which simplifies to approximately d × 0.00925. At 500 km (LEO), this gives 4.6 km. At 35,786 km (GEO), the same formula gives 331 km.
A GEO space mirror would deliver less than 0.02% of the sunlight intensity achievable from LEO — an immeasurable addition to natural illumination. The LEO–GEO tradeoff is fundamental: LEO gives intensity but only brief passes; GEO gives continuous coverage but useless intensity. There is no middle ground.
The Geometry Reflect Orbital Uses
If a space mirror must operate in LEO, the question becomes: which LEO orbit maximises the fraction of time during which the satellite is simultaneously in sunlight while the ground below it is in shadow? That window — the satellite lit, the ground dark — is the only time the mirror provides any contrast benefit.
The answer is a sun-synchronous orbit (SSO): a near-polar orbit with an inclination of approximately 97–98° that is engineered to precess (slowly rotate its orbital plane) at a rate matching Earth's annual orbit around the Sun. This keeps the orbital plane at a nearly constant angle relative to the Sun throughout the year. Crucially, a satellite in SSO always crosses any given latitude at approximately the same local solar time on every pass — making its behaviour predictable and consistent year-round.
Reflect Orbital's proposed constellation uses SSO specifically because it enables terminator-following geometry. At the terminator — the line separating Earth's sunlit and shadow hemispheres — the Sun is on the horizon from the perspective of the ground. A satellite in SSO can be phased so that it is flying near the terminator during its passes: in full sunlight at 500+ km altitude while the ground 1,000–2,000 km into the shadow side is in full darkness below it.
During these terminator passes — roughly at dawn and dusk local time — the satellite can reflect sunlight into the shadow zone. This is the operating envelope. Outside it, either the satellite is in Earth's shadow (not illuminated, useless as a mirror), or the ground below is in daylight (reflected sunlight adds nothing detectable against ambient solar illumination).
Why the Beam Is at Least 5 km Across
The ground footprint of a space mirror has a hard lower bound set by the angular size of the Sun. Because a mirror does not generate light — it reflects it — the reflected beam is an optical projection of the Sun's disc. The angular diameter of the Sun as seen from Earth is approximately 0.53° (31.6 arcminutes). This value varies slightly between perihelion and aphelion but is effectively constant for engineering purposes.
No optical system can produce a reflected beam narrower than the angular diameter of the source. A perfect mirror with zero surface roughness, perfectly flat, at 500 km altitude still produces a reflected beam with a minimum divergence angle of 0.53°. At 500 km altitude, a 0.53° cone has a base diameter of approximately 4.6 km. Real-world factors — mirror surface flatness, thermal distortion, attitude control error, atmospheric scintillation — widen the footprint beyond this minimum.
For Eärendil-1, which uses a relatively small mirror (approximately 75 m² sail area), the effective footprint is wider still because the mirror subtends a smaller solid angle than the Sun at the target distance, meaning the reflection is essentially a full projection of the solar disc. Larger mirrors approach the geometric minimum but never breach it. A 1-km² mirror at 500 km altitude would still produce a minimum 4.6 km footprint — the footprint is not a function of mirror area but of the Sun's angular diameter.
The practical consequence: a space mirror cannot be used as a precision illuminator for a small area. It is a wide-area diffuse light source, comparable to a large but dim floodlight — not a searchlight. See Solar Economics for what this means for energy delivery per square metre, and How Bright? for the ground illuminance calculations.
3 to 5 Minutes — The Hard Limit
A satellite at 500 km altitude moving at 7.6 km/s crosses its horizon-to-horizon arc in approximately 10–12 minutes from a ground observer's perspective. However, a space mirror can only deliver useful illumination during the portion of that arc when three conditions are simultaneously met: the satellite is in sunlight, the ground target is in shadow, and the satellite's orientation is correct to reflect sunlight toward the target.
The geometry collapses these conditions to a window of 3–5 minutes over any given ground point during the terminator passes at dawn and dusk. The satellite's angular velocity across the sky is highest when it is directly overhead (approximately 1° per second near zenith) and slows near the horizon — but near the horizon the atmospheric path length is greater and the mirror's projection angle is unfavourable. Peak illumination occurs near culmination, when the satellite is closest to the observer.
Two passes per day are usable for solar illumination (one at dawn, one at dusk, when the terminator geometry is right). The remaining 12–14 passes per day occur when the satellite is either in Earth's shadow (nighttime) or the ground below is lit by direct sunlight (daytime) — neither geometry permits useful mirror operation.
How Many Satellites for Continuous Coverage?
Providing near-continuous illumination over a single fixed ground location requires a constellation of satellites arranged to hand off coverage as each satellite exits the useful geometry window and the next enters it. The mathematics depend on orbital altitude, desired elevation minimum, and how precisely each satellite can be timed to the previous one's departure.
At 500 km altitude in a sun-synchronous orbit, a single satellite passes a given ground point twice per day with useful mirror geometry (once at dawn, once at dusk). Each pass lasts 3–5 minutes. To extend that to near-continuous coverage during the terminator window, you need satellites spaced such that when one exits the useful elevation angle, the next is already above the minimum elevation threshold.
This is why Reflect Orbital's business model begins with a single demonstrator satellite and a limited revenue case: Eärendil-1 demonstrates the hardware, the pointing system, and the energy delivery measurement. A full commercial product requires a constellation, which requires capital beyond a single-satellite demonstration. See Solar Economics for the economic analysis of what each satellite in a constellation can deliver per year to a ground solar farm.
Pointing — Harder Than It Looks
The reflective surface of a space mirror is mechanically simple — aluminised thin film, deployed to a large area. The attitude control system required to keep that surface pointed correctly is where most of the engineering complexity resides.
To reflect sunlight from orbit toward a specific ground target, the mirror must be oriented with an accuracy of a fraction of a degree. The required pointing precision is set by the footprint size requirement: if the target ground area is 5 km across and the satellite is at 500 km altitude, the allowable pointing error is approximately ±0.3°. This is achievable with modern reaction wheels and star trackers, but it must be maintained continuously while the satellite moves at 7.6 km/s relative to both the Sun and the ground target.
The pointing requirement is dynamic: the correct orientation changes continuously as the satellite moves along its orbit. The onboard attitude control system must compute and execute these orientation updates in real time, typically at rates of several cycles per second. Thin-film mirrors present additional challenges because large flexible surfaces can vibrate or oscillate — a mode called "structural dynamics" — that attitude control must damp without introducing its own oscillations.
Solar radiation pressure — the tiny force exerted by sunlight on the mirror's large surface area — also provides a continuous disturbance torque that the attitude control system must counteract. For large thin-film mirrors, this is a significant torque source. Correctly modelling and compensating for radiation pressure torque is a key engineering challenge that the Znamya experiments encountered in the 1990s and that Reflect Orbital's engineering programme must solve for Eärendil-1.