loading...
Effective (Roche) potential along the line of centres, in units of the separation
Radial-velocity curves of both stars (the observer sees the orbit at the chosen inclination)
How much of its Roche lobe each star fills, through one orbit (red line = overflow)
running...
How to use
Choose a preset (Alpha Centauri, Sirius, Algol, a cataclysmic variable...) or set masses, radii, separation (logarithmic slider), eccentricity and inclination. The yellow and blue teardrops are the Roche lobes; when a star's radius exceeds its lobe, gas streams through L1 (the Algol and cataclysmic-variable presets do this). Watch the third graph for eccentric orbits: the lobes breathe with the separation. The radial-velocity graph shows what a spectrograph would record. Space pauses, R resets.
The science
Two bodies, one ellipse. Two stars of mass M1, M2 orbit their common centre of mass. The relative orbit is a Kepler ellipse of semi-major axis a and eccentricity e, and each star follows a scaled copy of it: star 1 at -M2/(M1+M2) and star 2 at +M1/(M1+M2) of the separation vector. With a in AU, masses in solar masses and time in years, Kepler's third law is simply
P^2 = a^3 / (M1 + M2)
The position at time t comes from Kepler's equation M = E - e sin E with M = 2 pi t / P, then x = a (cos E - e), y = a sqrt(1-e^2) sin E, separation d = a (1 - e cos E). The stars move fastest at pericentre and slowest at apocentre (Kepler's second law).
Radial velocities. A spectrograph measures only the speed along the line of sight. For an orbit seen at inclination i (90 degrees = edge-on) the semi-amplitude of star 1 is
K1 = (M2 / (M1+M2)) (2 pi a sin i) / (P sqrt(1-e^2)), K2 = K1 M1 / M2, f(M) = P K1^3 (1-e^2)^(3/2) / (2 pi G) = M2^3 sin^3 i / (M1+M2)^2
The mass function f(M) is what a single-lined spectroscopic binary gives directly; it is a lower bound for the companion's mass.
Roche lobes. In the frame rotating with the stars (angular speed omega^2 = G (M1+M2) / d^3) a test particle feels the effective potential
Phi = - G M1 / r1 - G M2 / r2 - (1/2) omega^2 (x^2 + y^2)
Its stationary points are the five Lagrange points: L1 between the stars, L2 and L3 outside, L4 and L5 at the corners of equilateral triangles. The equipotential through L1 forms two touching teardrops, the Roche lobes. A star that grows past its lobe loses gas through L1 to its companion: this is how semi-detached systems such as Algol and cataclysmic variables work. The lobe radius is well fitted by Eggleton's formula
r_L / a = 0.49 q^(2/3) / ( 0.6 q^(2/3) + ln (1 + q^(1/3)) ), q = M(star) / M(other)
What the pictures show. The lobes are drawn for the instantaneous separation, so in an eccentric orbit they swell at apocentre and shrink at pericentre: a star can fill its lobe only near pericentre (see the third graph). The stream from L1 is a schematic arc, not a hydrodynamic solution.
Honest simplifications. (1) Point masses and spherical stars: no tides, no rotation, no mass loss or mass transfer feedback, no relativity. (2) The Roche geometry assumes a circular orbit and synchronous rotation and is applied quasi-statically at the instantaneous separation for eccentric orbits. (3) The overflow stream is illustrative. (4) Presets use rounded literature values (Alpha Centauri, Sirius, Algol, a generic cataclysmic variable) for orientation only. (5) Inclination only affects the radial velocities; the top-down picture is the true orbital plane.
Checks run in the page. L1 for equal masses against the midpoint; Kepler's third law from the period of an RK4 integration of the two-body equations; the integrated separation after 3.3 orbits against the solution of Kepler's equation; the integrator's energy drift; the mass function recovered from the simulated radial-velocity amplitude against its definition; and Eggleton's formula against the volume of the equipotential through L1 for q = 0.3 and q = 3.