Area products for multihorizon stationary black holes often have intriguing properties, and are often (though not always) independent of the mass of the black hole itself (depending only on various charges, angular momenta, and moduli). Such products are often formulated in terms of the areas of inner (Cauchy) horizons and outer (event) horizons, and sometimes include the effects of unphysical “virtual” horizons. But the conjectured mass independence sometimes fails. Specifically, for the Schwarzschild–de Sitter [Kottler] black hole in ($3+1$) dimensions it is shown by explicit exact calculation that the product of event horizon area and cosmological horizon area is not mass independent. (Including the effect of the third “virtual” horizon does not improve the situation.) Similarly, in the Reissner-Nordstrom–anti-de Sitter black hole in ($3+1$) dimensions the product of the inner (Cauchy) horizon area and event horizon area is calculated (perturbatively), and is shown to be not mass independent. That is, the mass independence of the product of physical horizon areas is not generic. In spherical symmetry, whenever the quasilocal mass $m(r)$ is a Laurent polynomial in aerial radius, $r=\sqrt{A/4\pi }$, there are significantly more complicated mass-independent quantities, the elementary symmetric polynomials built up from the complete set of horizon radii (physical and virtual). Sometimes it is possible to eliminate the unphysical virtual horizons, constructing combinations of physical horizon areas that are mass independent, but they tend to be considerably more complicated than the simple products and related constructions currently being mooted in the literature.

• Received 19 June 2012

DOI:https://doi.org/10.1103/PhysRevD.88.044014