If the rotor was mechanically strong enough to lift the desired weight
It wouldn't be. Because of the square-cube law, lift and mechanical strength grow with second power of linear size, but weight grows with third power. The result is that at RC-scale, everything has plenty of power and strength even if made sloppily (and thus cheaply) from common materials while human-carrying-scale is really stretching capabilities of the materials and engines we have.
A highly cambered airfoil would not be strong enough for full-scale helicopter blade.
is there any (aerodynamic) advantage to the solid aerofoil over the simpler curved surface?
I believe there is. No modern aircraft uses highly cambered airfoil.
Surely toy manufacturers could "close the bottom" of the rotor (equating to the aerofoil shape) even if it was hollow with neglible increase in weight.
No. Efficiency is not that important at RC scale and hollow plastic is much more difficult to make (remember, plastic is made by injecting it into a mold).
I understand the Bernoulli effect does not necessarily apply to the RC rotors
Airfoils is an airfoil whether it is flying straight or rotates. And since Bernoulli's principle is just conservation of energy for fluid flow, it always applies to it. It does not, however, mean concave lower surface would mean higher flow speed and lower lift, because the length of the path has nothing to do with lift. The air above the wing is faster—and reaches the trailing edge long before the air below—for completely different reasons. Highly cambered wings do have higher lift. They just have even higher drag.
it may be simply a Coanda effect
Coandă effect is about a fluid jet surrounded with still fluid around it, but there are no jets of air involved here, so it can't be Coandă effect.
but they do seem to lift well with just electric motors
That's the square-cube law again making the aerodynamic performance so much better relative to weight at RC scale.
I have seen discussion that there is a "scale up" effect but why and if so at what size does the simple rotor start to fail
It is not just scale up (square-cube law):
- Quad-copters with fixed pitch rotors (typical of the toy ones) are not controllable with failed engine. Human-carrying craft have to be.
- Fixed pitch rotors deal poorly with horizontal speed, because the lift asymmetry creates big loads. Large rotors need to lead-lag and flap to compensate for this.
Both making quadcopters impractical at large scales.
Usually, quad-copters are made up to low tens of kilograms maximum weight. But where exactly is the limit might be hard to tell since larger ones may not be made simply because there is not much use for them.