Mira Variables

A star that breathes once a year

Take a star like the Sun and run the clock forward about ten billion years. The core hydrogen is long gone; even the helium has been burned to carbon and oxygen. What remains is a tiny, electron-degenerate ash core no bigger than the Earth, wrapped in a colossal, tenuous envelope that has swollen to several hundred times the Sun's radius. The surface is so far from the energy source and so cool — around 3000 K — that it glows a deep, sullen red. This is an asymptotic-giant-branch star, and a large fraction of them are unstable. The envelope cannot sit still. It oscillates: expanding, cooling, falling back, reheating, on a cycle that takes roughly a year. That slow, enormous breathing is a Mira variable.

The defining feature, and the reason these stars were the first variables ever catalogued, is the sheer size of the brightness swing. The prototype, Omicron Ceti — "Mira", Latin for "the wonderful" — vanishes from naked-eye view entirely at minimum and returns to rival the brightest stars in its part of the sky at maximum. The visible flux changes by a factor of roughly a thousand. No other class of common, long-lived star does anything remotely like it. What makes the story interesting is that the star is not actually emitting a thousand times more energy at maximum — most of the swing is a trick of chemistry in the cooling gas.

Where Miras come from: the asymptotic giant branch

Mira variables are the visible signature of a specific, brief, and dramatic evolutionary phase. A star of initial mass between about 0.8 and 8 solar masses, having exhausted core helium, climbs the asymptotic giant branch (AGB). Its structure is now extreme: a degenerate carbon-oxygen core of about 0.5–0.7 M_☉, surrounded by a helium-burning shell, a hydrogen-burning shell above that, and an enormously extended convective envelope on top. During the thermally pulsing AGB (TP-AGB) phase the two shells take turns: the H shell builds up helium ash until it ignites in a runaway "thermal pulse" (helium-shell flash) every 10,000–100,000 years.

The numbers are worth holding onto. A typical Mira has:

Radius     R ≈ 200 – 500 R☉   (~1–2 astronomical units)
Luminosity L ≈ 2000 – 6000 L☉
Surface T  ≈ 2500 – 3500 K
Mass       M ≈ 0.8 – 2 M☉ (much of the original envelope already lost)
Mass loss  Ṁ ≈ 10⁻⁷ – 10⁻⁵ M☉/yr

If you placed a Mira where the Sun is, its photosphere would engulf the orbit of the Earth and reach toward Mars. The envelope is so distended and loosely bound that the star is shedding it: a slow, dusty wind carries material off into a circumstellar shell. The Mira phase lasts only of order 10⁵–10⁶ years before the envelope is entirely ejected, exposing the hot core as the central star of a planetary nebula and leaving behind a white dwarf. Miras are stars in the act of dying, and the pulsation is part of how they do it.

The engine: the kappa (opacity) mechanism

The pulsation is not random sloshing — it is a self-sustaining heat engine. The mechanism is the same opacity valve that drives Cepheids and RR Lyrae stars, but operating in a cool, hydrogen partial-ionisation zone instead of a helium one. The valve works like this:

1. Layer compresses  → hydrogen partially ionises → opacity κ RISES
2. High κ traps heat  → pressure builds            → layer pushed OUT
3. Layer expands, cools → hydrogen recombines       → opacity κ FALLS
4. Low κ lets heat leak → pressure drops             → layer falls BACK
   → repeat

For the valve to pump energy into the oscillation rather than damp it, the opacity must increase on compression — formally, the temperature exponent of the opacity must satisfy a condition first stated by Arthur Eddington and refined by Sergei Zhevakin. In a Mira the relevant zone is hydrogen ionisation at temperatures around 10⁴ K, which in such a cool, low-gravity envelope sits at a substantial depth and drives oscillations of huge amplitude. Crucially, Miras pulsate in the fundamental radial mode: the whole envelope moves in and out in phase, like a single spherical bellows, rather than in a higher overtone where nodes break the motion into shells. The fundamental mode has the largest radial excursion, which is exactly why Mira amplitudes dwarf those of overtone pulsators such as many semiregulars.

Why the visible swing is so enormous: TiO bands

Here is the part that surprises most people. The star's total energy output — the bolometric luminosity — changes by only about one magnitude over a cycle, a factor of roughly 2.5. Yet the visual magnitude changes by six, seven, even nine magnitudes. The extra factor of hundreds is not energy; it is chemistry.

As the pulsating photosphere cools toward minimum (down to ~2500–2900 K), the gas becomes cool enough for molecules to form. The dominant one for visible light is titanium oxide (TiO), with vanadium oxide (VO) adding to the effect. TiO has broad, strong absorption bands sitting squarely in the green-yellow part of the spectrum where the human eye and the standard V photometric filter are most sensitive. As the star cools, these bands deepen dramatically and devour visible photons, scattering the energy into the red and infrared. The Stefan–Boltzmann law alone says a 20% drop in temperature is only a factor of about 2.4 in total flux; the TiO bands turn that into a collapse of the V-band flux specifically.

The clean way to see this: measure the same star in the near-infrared K band (2.2 μm), far from the TiO bands. There, the amplitude shrinks to about one magnitude — close to the true bolometric variation. The thousand-fold visible swing is real, but it is a wavelength-selective illusion: most of the "missing" light at minimum has simply moved out of the band our eyes happen to use.

Bolometric amplitude:  ΔM_bol ≈ 1 mag   (factor ~2.5 in total power)
V-band amplitude:       ΔV     ≈ 6–8 mag  (factor ~250–1600)
K-band amplitude:       ΔK     ≈ 1 mag    (factor ~2.5)
Culprit:                temperature-sensitive TiO / VO absorption in V

Mira vs. its pulsating cousins, by the numbers

Mira variables are one branch of a family of radial pulsators. Comparing them across the relevant parameters makes the place of each class clear:

Class	Period	Type	Visual amplitude	Mode	Stage
Mira	100–1000 d	Cool red giant (M, S, C)	≥ 2.5 mag (often 6–8)	Fundamental	Thermally pulsing AGB
Semiregular (SRa/SRb)	30–1000 d	Red giant / supergiant	< 2.5 mag	Overtone / mixed	AGB / RGB
Classical Cepheid	1–100 d	F–K supergiant	0.1–2 mag	Fundamental / 1st overtone	Helium-burning loop
RR Lyrae	0.2–1 d	A–F horizontal branch	0.3–1.5 mag	Fundamental / overtone	Core helium burning
Delta Scuti	0.02–0.3 d	A–F dwarf / subgiant	< 0.9 mag	Many (p-modes)	Main sequence / subgiant
RV Tauri	30–150 d	F–K supergiant	up to 3–4 mag	Alternating deep/shallow	Post-AGB

Two things stand out. First, Mira amplitudes are extreme precisely because they pulsate in the fundamental mode in a very cool, very extended envelope where TiO amplification is available — a combination no other class has. Second, the periods are long because the dynamical timescale of a star scales as the inverse square root of mean density, and a Mira's mean density is absurdly low: spreading two solar masses over a radius of 1–2 AU gives a mean density well below that of air.

Standard candles in the infrared

Because pulsation period is set by the star's mean density and structure, and luminosity tracks core mass, Mira variables obey a period-luminosity relation — the same physical idea that makes Cepheids cosmic yardsticks. For oxygen-rich (M-type) Miras the relation in the near-infrared K band is approximately

M_K ≈ −3.5 log₁₀(P / days) + constant      (scatter ≈ 0.13–0.15 mag)

A 300-day oxygen-rich Mira sits near M_K ≈ −7.5; a 400-day one is brighter still. The infrared is essential: the TiO bands and circumstellar dust that wreck the visible-band amplitude matter far less at 2.2 μm, so the K-band light curve cleanly samples the bolometric pulsation. With absolute magnitudes near −7 to −8, Miras are bright enough to be seen across the Local Group, and they are far more numerous than Cepheids of comparable luminosity. They trace an intermediate-age-to-old population, complementing the young Cepheids, and have become a key independent rung of the distance ladder — used to calibrate distances to the Large Magellanic Cloud and, with JWST and adaptive-optics infrared imaging, to cross-check the Cepheid scale that anchors the Hubble tension debate.

Dust, mass loss, and a thirteen-light-year tail

The pulsation does more than make the star blink — it is the trigger for the most important thing a Mira does for the galaxy: it returns processed matter to interstellar space. Each cycle, the shock from the bellows-like pulsation levitates gas to a few stellar radii where it is cool enough (below ~1500 K) for dust grains — silicates around oxygen-rich Miras, amorphous carbon around carbon-rich ones — to condense. Starlight then pushes on the grains via radiation pressure, the grains drag the gas along by collisions, and the result is a dust-driven wind carrying 10⁻⁷ to 10⁻⁵ M_☉ per year off the star at 5–30 km/s.

Over the Mira lifetime this strips away most of the envelope, building a circumstellar shell rich in molecules (often detected as SiO, H₂O, and OH maser emission — the "OH/IR stars" are dust-enshrouded extreme Miras). This is a primary channel by which carbon, nitrogen, and s-process elements forged in the AGB interior — see s-process nucleosynthesis — are seeded into the next generation of stars and planets.

Mira itself supplied a spectacular demonstration. The star moves through the interstellar medium at about 130 km/s, and in 2007 NASA's GALEX ultraviolet satellite revealed that its shed material trails behind it in a turbulent tail roughly 13 light-years long, with a bow shock at the leading edge — a fossil record of about 30,000 years of mass loss. No other star was known to wear such a comet-like wake.

Mira itself and other famous examples

• Omicron Ceti (Mira A). The prototype, in Cetus, about 300 light-years away. Period ≈ 332 days; it ranges from about visual magnitude 3.5 at maximum (occasionally brighter than 2) down to about 9–10 at minimum — a swing of more than 6 magnitudes. Its variability was recorded by David Fabricius in 1596 and recognised as periodic by Johannes Holwarda in 1638, making it the first periodic variable known. It has a white-dwarf companion (Mira B) accreting from its wind.
• Chi Cygni. An S-type Mira in Cygnus with a long 408-day period and one of the largest amplitudes known — up to about 11 visual magnitudes between extremes (magnitude ~3.3 at best maximum to ~14 at minimum).
• R Leonis. A bright, well-studied oxygen-rich Mira in Leo, period ≈ 312 days, magnitude ~4.4 to ~11.
• R Hydrae. Famous for a period that has measurably shortened over centuries — from about 500 days in the 1700s to about 385 today — interpreted as a response to a recent thermal pulse, a rare case of stellar evolution visible on a human timescale.
• IK Tauri and OH/IR stars. Extreme, dust-enshrouded Miras whose envelopes are so thick the star is invisible in the optical and only emerges in the infrared and as OH/SiO masers — the high-mass-loss end of the sequence, on the verge of becoming planetary nebulae.

Common misconceptions and edge cases

• "A Mira gets a thousand times more luminous." No. The total energy output changes by only a factor of about 2.5 (one magnitude). The thousand-fold figure is the visible-band flux, inflated by temperature-sensitive TiO absorption. Measure in the infrared and the swing nearly vanishes.
• "Miras are pulsating supergiants like Betelgeuse." They are not. Miras are low- and intermediate-mass AGB giants (initial mass < 8 M_☉), not massive supergiants. Betelgeuse is a semiregular red supergiant of ~15–20 M_☉ destined for core collapse; a Mira will end as a quiet white dwarf inside a planetary nebula.
• "The period is perfectly clockwork." Mira periods are stable enough to define the class but drift cycle-to-cycle by a few percent, and the maxima vary in brightness from one cycle to the next. A handful (R Hya, R Aql, T UMi) show secular period changes traceable to recent helium-shell thermal pulses.
• "All Miras are M-type (oxygen-rich)." Most are, but the class also includes S-type (carbon/oxygen ≈ 1, with ZrO bands) and C-type carbon Miras, depending on how much dredged-up carbon the AGB star has accumulated — see carbon stars. Carbon and oxygen Miras follow slightly different period-luminosity relations and form different kinds of dust.
• "Radius and brightness peak together." They do not. Because of the way the pulsation shock and temperature lead or lag the radius change, visual maximum tends to occur near the phase of fastest expansion (highest temperature), not at maximum radius. The light curve and radial-velocity curve are out of phase — a classic diagnostic of the underlying pulsation physics.