Binary Stars
Superhumps: The Slowly Precessing Eccentric Disk in Dwarf Novae
Roughly every few weeks a small binary star brightens by a factor of a hundred, and buried in that flare is a clock that ticks about 3% slower than the stars' orbit — a photometric wobble called a superhump that repeats every few hours and slowly drifts in phase. That extra beat is not the orbit at all: it is the whole accretion disk, made lopsided and eccentric by tidal forces, wheeling around the binary once every few days.
Superhumps are periodic brightness humps seen during the superoutbursts of SU UMa-type dwarf novae (and related cataclysmic variables), with a period a few percent longer than the orbital period. They are the observational signature of a tidally driven eccentric, prograde-precessing accretion disk — a resonance phenomenon that lets astronomers weigh binaries they can barely resolve.
- TypePhotometric period in cataclysmic variables
- RegimeSU UMa dwarf novae, q ≲ 0.25–0.33
- Discovered / explainedVogt & Warner 1970s; Whitehurst 1988, Osaki 1989
- Typical period excessε = (P_sh − P_orb)/P_orb ≈ 1–7%
- Key relationε ≈ 0.18 q + 0.29 q² (Patterson 2005)
- Observed inVW Hyi, Z Cha, OY Car, WZ Sge, ER UMa, AM CVn stars
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What a superhump actually is
A cataclysmic variable is a close binary in which a white dwarf accretes hydrogen-rich gas from a low-mass, Roche-lobe-filling companion via an accretion disk. In dwarf novae, thermal instabilities in that disk trigger outbursts of 2–5 magnitudes every few weeks. A subclass, the SU UMa stars (all with short orbital periods, P_orb ≲ 2 h, below the CV "period gap"), occasionally show a longer, brighter superoutburst.
During a superoutburst a smooth, tooth-shaped brightness modulation appears in the light curve. Its period, the superhump period P_sh, is typically a few percent longer than the spectroscopically measured orbital period. Crucially, P_sh is not a beat you can build from the orbit alone — it is the beat between the orbital motion and a slow precession of a distorted disk. The fractional difference,
- ε = (P_sh − P_orb) / P_orb, the "superhump period excess,"
is the single most useful number: it encodes the binary mass ratio q = M_donor/M_wd.
The mechanism: 3:1 resonance and a lopsided, precessing disk
The physics was pinned down by R. Whitehurst (1988), who found in 2D hydrodynamic simulations that a disk can go tidally unstable and eccentric, and by Y. Osaki (1989), who tied it to dwarf-nova superoutbursts in the thermal–tidal instability (TTI) model. The key ingredient is a 3:1 orbital resonance: gas orbiting the white dwarf at a radius where its orbital frequency is exactly three times the binary frequency receives a coherent tidal kick from the donor every orbit.
That resonance pumps eccentricity into the outer disk, turning a circle into a slowly rotating ellipse. Because the tidal potential is fixed to the binary, the eccentric disk undergoes prograde apsidal precession — its long axis wheels around once every precession period P_prec (a few days), much longer than P_orb. As the elongated disk sweeps around, the tidal stressing and viscous dissipation peak once per beat between orbit and precession:
- 1/P_sh = 1/P_orb − 1/P_prec, so P_sh > P_orb.
The resonance only fits inside the tidal (Roche) radius when the donor is light, i.e. q ≲ 0.25–0.33 — which is why superhumps mark short-period systems.
Key numbers and a worked example
The precession rate of a test particle at the 3:1 resonance depends almost entirely on q, so ε maps cleanly onto mass ratio. The workhorse empirical relation from Patterson et al. (2005), calibrated on eclipsing CVs with independently known q, is:
- ε ≈ 0.18 q + 0.29 q²
- Kato et al. (2009) update: ε ≈ 0.16 q + 0.25 q² (using "stage A" superhumps that reflect the pure dynamical rate)
- Knigge (2006): q ≈ 0.114 + 3.97 (ε − 0.025)
Worked example. Take Z Chamaeleontis: P_orb ≈ 0.0745 d (1.79 h) and P_sh ≈ 0.0772 d. Then ε = (0.0772 − 0.0745)/0.0745 ≈ 0.036. Solving 0.18q + 0.29q² = 0.036 gives q ≈ 0.18. With a ~0.85 M_sun white dwarf this implies a donor of ~0.15 M_sun — consistent with the eclipse-derived value. Typical SU UMa numbers: P_orb 1.3–2.2 h, ε 0.02–0.07, P_prec 2–5 d, superhump amplitude 0.1–0.4 mag.
How superhumps are observed and dissected
Superhumps are found by high-cadence time-series CCD photometry through a superoutburst, followed by Fourier / phase-dispersion analysis; the orbital period comes independently from radial-velocity spectroscopy or (best) eclipse timing. The global VSNet campaigns led by Taichi Kato have measured superhumps in hundreds of systems, and space photometry from Kepler/K2 and TESS now resolves their evolution in exquisite detail.
A hallmark is that P_sh is not constant. Kato's surveys classify three stages:
- Stage A — early, longest period; reflects the pure dynamical precession at the 3:1 radius (best for q).
- Stage B — period systematically decreasing, a pressure (gas-dynamical) effect as the disk shrinks; often shows a positive dP/dt in short-P systems.
- Stage C — a later, roughly constant shorter period.
This stage evolution, plus "early superhumps" and "late superhumps," turns a single light curve into a diagnostic of disk radius, viscosity, and the resonance's grip.
Relatives: negative superhumps, early superhumps, and permanent superhumps
Positive (ordinary) superhumps are the eccentric-disk case described above. They have close cousins that are easy to confuse:
- Negative superhumps: period a few percent shorter than orbital. Traditionally attributed to retrograde nodal precession of a disk tilted out of the orbital plane, so the bright spot / hot inner disk is periodically better exposed; the origin of the tilt remains debated, and recent work argues some arise from retrograde apsidal precession of a cool eccentric disk.
- Early (double-wave) superhumps: seen in the first days of extreme WZ Sge superoutbursts; nearly at P_orb, driven by the deeper 2:1 resonance (needs q ≲ 0.08) that produces a two-armed spiral pattern rather than an eccentric ring.
- Permanent superhumps: in nova-like CVs and AM CVn stars with high, steady accretion, the disk stays large enough that the eccentric mode never switches off.
The distinguishing test is simply the sign of ε and whether the modulation persists between outbursts.
Why it matters, and what's still open
Superhumps are the most reliable way to estimate mass ratios of unresolved, non-eclipsing CVs, feeding directly into CV evolution: how donors shrink, cross the period gap, and reach the period minimum (~80 min) before becoming "period bouncers" with substellar donors. The ε–q relation extends the same trick to the ultracompact AM CVn helium binaries and has even been invoked for a 693.5 s modulation in the X-ray binary 4U 1820−30.
Open questions remain sharp:
- The ε–q calibration is anchored by assuming ε → 0 as q → 0, which is not measured; the low-q end is poorly constrained, adding systematic error to bouncer masses.
- The physical origin and persistence of the disk tilt behind negative superhumps is unresolved.
- Whether "stage" period changes trace disk-radius shrinkage, pressure effects, or eccentricity growth is still modeled with SPH simulations.
Famous laboratories — VW Hyi, OY Car, Z Cha, WZ Sge, ER UMa — remain the benchmarks against which every superhump model is tested.
| Property | Positive (ordinary) superhump | Negative superhump | Early superhump |
|---|---|---|---|
| Period vs. orbit | Longer: P_sh ≈ 1.01–1.07 P_orb | Shorter: P_nsh ≈ 0.95–0.98 P_orb | ≈ P_orb (within ~0.1%) |
| Underlying motion | Prograde apsidal precession of eccentric disk | Retrograde nodal precession of a tilted disk | 2:1 resonance / disk overflow shadow |
| Driving resonance | 3:1 orbital resonance | Disk tilt out of orbital plane | 2:1 resonance (needs q ≲ 0.08) |
| Excess ε | +0.01 to +0.07 | −0.02 to −0.05 | ≈ 0 |
| When seen | Body of the superoutburst (SU UMa stars) | Some standstills/superoutbursts; VUV/nova-likes | First days of WZ Sge superoutbursts |
Frequently asked questions
Why is the superhump period longer than the orbital period?
The superhump is the beat between the orbital motion and the slow prograde precession of the eccentric disk: 1/P_sh = 1/P_orb − 1/P_prec. Because the disk's long axis wheels forward once every few days in the same sense as the orbit, the pattern of tidal stressing recurs slightly slower than once per orbit, making P_sh a few percent longer than P_orb. This positive excess is the defining feature of ordinary (positive) superhumps.
What is the 3:1 resonance and why does it need a low mass ratio?
It is the radius where disk gas orbits the white dwarf three times per binary orbit, so the donor's tidal tug arrives in phase each cycle and pumps eccentricity into the disk. That resonance radius only fits inside the disk's tidal truncation radius when the donor is much lighter than the white dwarf, i.e. mass ratio q = M_donor/M_wd ≲ 0.25–0.33. Above that limit the resonance lies outside the disk and superhumps do not form.
How do astronomers get the mass ratio from a superhump?
They measure the superhump period excess ε = (P_sh − P_orb)/P_orb and invert an empirical relation. The most-used is Patterson et al. (2005), ε ≈ 0.18q + 0.29q²; Kato et al. (2009) and Knigge (2006) give refined versions using the early 'stage A' superhump, which best reflects the pure dynamical precession rate. The method works even for faint, non-eclipsing binaries that cannot be weighed any other way.
What is the difference between a superhump and a normal dwarf-nova outburst?
A normal outburst is a thermal disk instability that brightens the system for a few days; it has no periodic hump. A superoutburst is longer and brighter, and only during it does the disk expand past the 3:1 resonance, become eccentric, and precess — producing the superhump modulation. So the superhump rides on top of the superoutburst and reports on disk shape, not on the outburst trigger itself.
What are negative superhumps?
Negative superhumps are modulations with periods a few percent shorter than the orbital period. They are usually explained by retrograde nodal precession of an accretion disk tilted out of the orbital plane, which changes how much of the hot inner disk we see each cycle. The origin of the tilt is not fully understood, and some recent models instead invoke retrograde apsidal precession of a cool eccentric disk.
Who discovered and explained superhumps?
Superhumps were noticed photometrically in SU UMa stars in the 1970s (work by Vogt, Warner and others). The eccentric-disk mechanism was established by Ray Whitehurst's 1988 hydrodynamic simulations showing tidal instability, and Yoji Osaki's 1989 thermal–tidal instability model linked it to superoutbursts. Large modern surveys by Taichi Kato and space photometry from Kepler and TESS have since mapped their detailed behavior.