Criteria for IP Classification

On this site, I use a 5-tier system to classify IPs and IP candidates from "ironclad" to "doubtful," based on my judgement on the strength of published evidence for and against their IP credentials.

Requirement to be classified as an IP

As of 2026 August, here are my criteria for considering an object to be a confirmed IP.

There is at least one published optical spectrum consistent with a CV classification: a blue non-stellar continuum plus emission lines of hydrogen Balmer series and helium lines, particulary HeII 4686. Other clues (X-ray and optical luminosities, optical colors, presence of multiple periods in the X-rays and/or the optical) can point to an object being an IP. However, I am reluctant to consider the case closed without at least one optical spectrum.
The orbital period is known securely or at least inferred, and is within the range consistent with a CV classification. Uncertain values (including possible 1-day aliases) are acceptable, but I now exclude systems in which the orbital period is completely unknown. There are "Probable" IPs without orbital periods, some of which were previously considered "confirmed." This new exclusion because the orbital period is such a fundamental property of any CVs, and an IPs without an orbital perioda cannot be put on the Porb-Pspin diagram, for example.
The presence of a persistent and coherent period shorter than the orbital period is established. This is usually the spin period, but it could also be the sideband period.

That is, the object in question must be a CV (criteria 1 and 2 above), except that I do allow the orbital period to be estimated from the frequency difference between two stable periods. The 3rd criterion means that the spin of the white dwarf manifests itself in the data - the expression "persistent and coherent" hopefully eliminates CVs hosting pulsating white dwarfs.

The 5 Star System

"Ironclad" is used for systems that so obviously belong to the class that there are no arguments about the IP nature (e.g., AO Psc). This typically requires multiple independent confirmation of key properties.
"Confirmed" indicates that the system in question may have a key parameter that was measured only by one group, or that one or more of the periods may not be known accurately.
"Probable" means there is some evidence in favor of the IP nature and no evidence strongly against it.
"Possible" often means too little is known about the system. <>I>"Doubtful" means that there is at least one piece of evidence against an IP classification. My philosophy for this catalog is to never remove an object if it was once proposed to be an IP in refereed journal, even if there is a new consensus that it is not (including objects that are not even CVs) - so these will receive the designation of "doubtful."

Comparison with Patterson (1994)

Joe's 1994 review contains a section titled "The Cast: Rules of the Audition." In my mind, these rules apply to the subclassification of confirmed CVs as IPs.

Of his 6 criteria, I would not be bothered by a lack of circular polarization or a relatively soft X-ray spectrum - there are confirmed IPs with a soft X-ray spectrum and/or with stringent upper limit on the circular poralization. On the other hand, I place a heavier emphasis on the emission line pulsation. Penning's pioneering work as well as many papers by Coel Hellier and company show that IPs have radial velocity (or its proxy, V/R ratio) modulation on the spin period. This is easy to underatsnd: a significant fraction of optical emission lines in IPs come from the "accretion curtain." So, if time-resolved optical spectroscopy with a sufficient sensitivity fails to show such a behavior, I would take that as a strong negative.

Of course, the usual criteria for IP classification is optical and X-ray spin modulations. When the classification depends on the optical periodicity, it should be coherent and persistent to a high degree. It has to be repeatedly observed, ideally over several years; If you see apparent modulations for a few cycles --- don't get too excited. I think every CV observer should read Warner's Warning (or see th PDF version of the same) before attempting to identify periodicities in CVs. If I were writing my papers on KO Vel now, with the same data, I would probably write a very different paper.

When X-ray pulses are involved, I would relax the requirement --- two independent X-ray pulse period determinations would be sufficient for me (assuming it is known to be a CV and the periods agree). If it depends on one X-ray observation, the data duration must be sufficient and the periodicity must stand out clearly above the noise. Note that methods designed to detect the weakest of X-ray period are also susceptible to false positives.

I would also happily accept somewhat weak optical detection (in terms of establishing persistence and coherence) combined with a single X-ray detection of the same period.

Koji Mukai, originally written in 2002 February, updated in 2014 September, the link to Warner's Warning updated in 2022 January. Substantially updated in 2026 August.

Laast updated 2026 August 17

Please send your comments, suggestions etc. to Koji.Mukai@nasa.gov and/or Koji.Mukai@umbc.edu