There's no magic number, I think. If you're reading the Sandia test reports like I did, you'll figure that "square" is a good starting point, then make something, and adjust it from there. One thing that stopped me working with VAWT's was that increasing the rotor diameter drastically reduced the RPM's (as you say in #2). Even running without a load I found that it was turning very slowly to begin with. Considering making it taller, but I could tell that the oscillations of the top, even at 4' tall, would get worse unless I put a bearing and supported it with wires at the top. And that was that.
Your questions #2 and #3 both lead to the "solidity" of the VAWT rotor, whatever its design. A savonius has nearly 100% solidity, and a darrieus may have about 20% or so. Increasing solidity, by adding a new blade, or widening the chord of the existing ones, for example, will slow it down. Same holds for HAWT's for that matter.
To brake a VAWT with a PM motor you would "short" the motor's wires. Connect the leads directly to each other, ensuring that you do this on the side of the rectifier / diode that gives the battery protection. This certainly works with HAWT blades, and the VAWT's don't have that much more torque that you can't use this method.
Of course, if you insist on breaking the motor, a 6-foot drop to a concrete floor would do...
