So as far as I understand, at 100km atmosphere exists, although very faint. Could someone approximate the speed of a sphere at 50km altitude after re-entry from a high orbit? It might sound as a wierd question but I am trying to calculate a factor and need an approximation of speed that would be realistic. An answer like 'hypersonic' would suffice with a minimum realistic value (educational guess). I am satisfying my curiosity of approximating a possibility of something that's quite nuts.
tl;dr: At 50 km altitude, about Mach 8-10 from LEO, and about Mach 15 returning from the Moon (which is coming in roughly the same speed it would be from a high Earth orbit). Your milage will vary depending on a lot of aerodynamic details!
See the question How does a Reentry Breakup Recorder survive reentry and then broadcast its data before impact? for the first image of the actual spacecraft, and then @Uwe's excellent answer for a lot more details on the spacecraft.
For the data shown below, see the 2013 presentation by Andrew S. Feistel, Michael A. Weaver, and William H. Ailor, of the Vehicle Systems Division, The Aerospace Corporation: Comparison of Reentry Breakup Measurements for Three Atmospheric Reentries at the 6th IAASS Conference: Safety is Not an Option, May 2013.
So for a small blunt reentry craft deorbited from LEO, it looks like about Mach 8 to 10.
below: From page 10 of Comparison of Reentry Breakup Measurements for Three Atmospheric Reentries
From NASA Technical Note TN D-6792: The Aerodynamic Environment of the Apollo Command Module during Superorbital Entry by Dorothy B. Lee and Winston D. Goodrich (1972) one of the examples shown gives about 16,000 feet/sec or about 4850 m/s or about mach 15 when re-entering the atmosphere from cis-ulnar space or "Superorbital" speed.
Atmospheric effects during reentry usually start around 120km.
The atmosphere extends far beyond 100km. Even at 400km the ISS is constantly slowing down, enough to require periodic reboosts.
The exosphere extends much further again, up to 10,000km were it merges with the solar wind. The exosphere will also slow spacecraft down, but the amount only becomes significant over longer timescales. https://en.m.wikipedia.org/wiki/Exosphere
There's not an answer to that. At least not precisely. The compromises are:
Slowing down higher up usually requires lift generation or very high drag to mass. If you're in a space shuttle this is a lot easier than in a capsule or just a chunk or rock. You also generate a bit of heat over a significant time. Favouring not ablative heat-shielding.
Slowing down lower down requires surviving greater aero-dynamic stresses. It also means higher temperatures, but they must be endured for less time. This favours ablative heat- shielding.
The faster you are going to start with, out scrubbing as much speed as you can in the upper atmosphere will still leave you a fair amount to get rid of lower down.
This graph: https://www.semanticscholar.org/paper/Radiation-Ablation-Coupling-for-Capsule-Reentry-via-Leyland-Morgan/471eb8a136f41006def99b65e8ccf3cc5592322e/figure/7 show this in action for Saturn V capsules vs the Space shuttle. These are pretty much opposite ends of the spectrum on all three counts and as predicted the capsule does most of its slowing down at about half the altitude of the shuttle. I think these two should be good bounds though. You would need quite exotic designs to considerably extend the "slowing down" altitude out of those ranges.