Aims. The orbit of the outer satellite Alexhelios of (216) Kleopatra is already constrained by adaptive-optics astrometry obtained with the VLT/SPHERE instrument.
However, there is also a preceding occultation event in 1980 attributed to this satellite. Here, we try to link all observations, spanning 1980-2018, because the nominal orbit exhibits an unexplained shift by +60 degrees in the true longitude.
Methods. Using both a periodogram analysis and an l = 10 multipole model suitable for the motion of mutually interacting moons about the irregular body, we confirmed that it is not possible to adjust the respective osculating period P-2.
Instead, we were forced to use a model with tidal dissipation (and increasing orbital periods) to explain the shift. We also analysed light curves spanning 1977-2021, and searched for the expected spin deceleration of Kleopatra.
Results. According to our best-fit model, the observed period rate is (P) over dot(2) = (1.8 +/- 0.1) x 10(-8) d d(-1) and the corresponding time-lag Delta t(2) = 42 s of tides, for the assumed value of the Love number k(2) = 0.3.
This is the first detection of tidal evolution for moons orbiting 100 km asteroids. The corresponding dissipation factor Q is comparable with that of other terrestrial bodies, albeit at a higher loading frequency 2 vertical bar omega - n vertical bar.
We also predict a secular evolution of the inner moon, (P) over dot(1) = 5.0 x 10(-8), as well as a spin deceleration of Kleopatra, (P) over dot(0) = 1.9 x 10(-12). In alternative models, with moons captured in the 3:2 mean-motion resonance or more massive moons, the respective values of Delta t(2) are a factor of between two and three lower.
Future astrometric observations using direct imaging or occultations should allow us to distinguish between these models, which is important for our understanding of the internal structure and mechanical properties of (216) Kleopatra.