1. Current plans do not appear to have a carrier-borne AAR asset for F35B, just as was case for SHAR/GR9. However, F35B combat radius is greater than SHAR/GR7 and the PnD fit (compare with SHAR) less likely to incur operational restrictions.
2.
AFAIK, the initial consideration of "steam" cats was limited to digging out the old BS6 drawings and comparing them to the available space. The team later got limited access to US C13 cat info to update weight and space provision, but these are subject to ITAR issues (as is EMALS for that matter). I don't know whether specific boiler fits were actually looked at. I'm pretty sure costing was unlikely to have been detailed enough to say whether cheaper than EMALS or not. Likely to have been cheaper only in an initial sense in any case. TL support for steam systems that would be unique in the fleet outside the SSN/SSBN community and requiring a significant number of below decks bodies to support would not have been cheap. Nor would finding them accommodation in the ship.
3. The type of boiler was probably not looked at in detail (see above). I'd bet the ranch no-one ever considered using steam turbines for propulsion. The system would have used steam reservoirs on 2 deck (similar to the US systems) charged off a dieso-fired set of boilers. That also provides a level of persistence / capacitance. Without doing the maths, you're looking at a boiler of a few tens of tons, not a drama in a ship that size. As with all steam systems, in operation you have to keep them warm, so it's unilkely that the plant would ever be shut down completely.
4. The designs have NEVER considered any use of propulsion steam turbine. The early concepts used 4 off WR21 (thank God THAT never got any further)driving electric motors , before they got a bit bigger and ended up as a GT/diesel combo for the generating plant.
5. No - EMALS is not self-contained wrt power generation. It interfaces with the CVN electrical distribution system, via a power controller. This may have been one of the issues with the UK fit. We will distribute shipwide at 11kVA (standard offshore practice). The US distribution system is at a lower kVA rating.
6. Three navies operate steam cats - Brazil, France and the US. All use steam powered carriers.
7. The CVF has always been planned to operate as GT (or primarily GT) powered ship. CTOL operations have always been seen as a "fallback" option in case STOVL failed. It was only when F35B started going seriously wrong that CTOL ops came to the fore. The reasons are complex and include political and technical factors.
On the political side, STOVL was always attractive to the RN as it was what they had become used to. It is also relatively cheap in terms of people and if the RN has failed in the CVF programme, it is in continually being unprepared to pay for and defend the f/w FAA until it was too late. Once FCBA had been captured and became JCA, the RAF liked STOVL because (they thought) it meant they could "replace" the Harrier GR capability and thus maintain squadron badges and aircrew numbers, without going to sea regularly, as would be required if they were the augment to a CTOL CAG. The pollies liked it because Rolls Royce kept a foothold in US military engine supply.
The technical side is a bit more straightforward. The timeline for gestation of the ship design has unfortunately spanned a period where everybody was CERTAIN that steam aboard ships is a bad thing (probably rightly) because it's difficult, hard to maintain, has safety issues and eats manpower. At the same time, the "alternative" EM technology was not sufficiently mature to include in the design without an eye-watering risk allocation (cost!). It is only comparatively recently that the EM technology has been considered low enough risk to be affordable. Unfortunately, by that time the build of the ship has progressed, making fitting it more expensive - although I certainly don't believe the £2Bn figure quoted. If we were at the same stage now as we were in 2002, then I have no doubt whatsoever that a more detailed CTOL option would have been prepared as part of the adaptable design and that it might even have won out in a cost-benefit analysis. However, we're not and it didn't - for whatever reason.