Context. Mergers of two stellar-origin black holes are a prime source of gravitational waves and are under intensive investigation. One crucial ingredient in their modeling has been neglected: pair-instability pulsation supernovae with associated severe mass loss may suppress the formation of massive black holes, decreasing black-hole-merger rates for the highest black-hole masses. Aims. We demonstrate the effects of pair-instability pulsation supernovae on merger rate and mass using populations of double black-hole binaries formed through the isolated binary classical evolution channel. Methods. The mass loss from pair-instability pulsation supernova is estimated based on existing hydrodynamical calculations. This mass loss is incorporated into the StarTrack population synthesis code. StarTrack is used to generate double black-hole populations with and without pair-instability pulsation supernova mass loss. Results. The mass loss associated with pair-instability pulsation supernovae limits the Population I/II stellar-origin black-hole mass to 50 M⊙, in tension with earlier predictions that the maximum black-hole mass could be as high as 100 M⊙. In our model, neutron stars form with mass 1−2 M⊙. We then encounter the first mass gap at 2−5 M⊙ with the compact object absence due to rapid supernova explosions, followed by the formation of black holes with mass 5−50 M⊙, with a second mass gap at 50−135 M⊙ created by pair-instability pulsation supernovae and by pair-instability supernovae. Finally, black holes with masses above 135 M⊙ may potentially form to arbitrarily high mass limited only by the extent of the initial mass function and the strength of stellar winds. Suppression of double black-hole-merger rates by pair-instability pulsation supernovae is negligible for our evolutionary channel. Our standard evolutionary model, with the inclusion of pair-instability pulsation supernovae and pair-instability supernovae, is fully consistent with the Laser Interferometric Gravitational-wave Observatory (LIGO) observations of black-hole mergers: GW150914, GW151226, and LVT151012. The LIGO results are inconsistent with high (≳ 400 km s-1) black hole (BH) natal kicks. We predict the detection of several, and up to as many as ~60, BH-BH mergers with a total mass of 10−150 M⊙ (most likely range: 20−80 M⊙) in the forthcoming ~60 effective days of the LIGO O2 observations, assuming the detectors reach the optimistic target O2 sensitivity.