If we allow that the energy to allow quantum tunnelling through a potential barrier has been borrowed by the Heisenberg energy-time uncertainty relation, from the vacuum energy. How is the borrowed energy paid back?
If we allow that the energy to allow quantum tunnelling through a potential barrier has been borrowed by the Heisenberg energy-time uncertainty relation, from the vacuum energy
... we don't. That is a simplified picture (more accurately, a lie-to-children), told at an introductory and popular-science level to explain how quantum tunnelling works without going into the full workings of quantum wave mechanics, and which does not represent the modern understanding of tunnelling at any substantial level of accuracy.
As such, any perceived inconsistencies within that picture are irrelevant - the picture is useless to begin with.
As to how which picture does hold for tunnelling, it is basically the wave mechanics of evanescent waves: wave amplitudes cannot go discontinuously from nonzero to zero, and if you try to stop a wave (via, say, total internal reflection) there will be some amount of 'leakage' of the wave amplitude into the 'forbidden' region with an exponential decay as you go into that region. If that region is finite, then that evanescent wave can couple to a propagating mode on the other side and be on its way. And if that sounds like the particle picture with energies &c kind of got lost - then welcome to the world of wave-particle duality.