Speculation about Mars missions produces an irresistable temptation to design paper spaceships, and I won't even try to resist. So here we go:
My bias, as expressed last post, is for reaching Mars and returning on fast transfer orbits, making the one way trip in approximately three months. Allowing a few weeks on (or at least orbiting) Mars, the mission can be done in six months and change.
The alternative is using Hohmann transfer orbits, more or less, and accepting an 18-month round trip. This can be done with chemfuel rockets, and broadly speaking we already know how to do that part. What we don't know is how to send humans into space for 18 months and get them back in good health.
Six months corresponds to the currently accepted mission duration for the ISS. Going much longer will be much harder on the crew. On the other hand, the 'fast' six-month mission calls for an electric drive; even nuclear thermal comes up short.
An electric drive with sufficient performance is semi-speculative. The Dawn probe has three (redundant) ion thrusters with a combined mass of 129 kg, thus 46 kg each, while its solar wings come in at 204 kg. Each thruster delivers 92 mN of thrust from 2600 watts of electric power.
The combined mass of one thruster plus solar wings thus comes to 247 kg, just about 0.01 kW/kg. We need power performance about 100 times better, approaching the figure of merit I have often mentioned here, 1 kW/kg is the standard figure of merit.
The reason for using this figure of merit shows up in sims done on the Rocketpunk Manifesto TravelPlanner. Specifically I looked at a baseline vehicle capable of putting on 29 km/s of delta v in 60 days of acceleration, allowing a 30-day coasting period. This is more efficient than a classical brachistochrone orbit, which expends energy and propellant on putting on speed, then (literally!) turning around and taking it back off again.
Exhaust velocity is a little over 30 km/s, giving a mass ratio of 2.5: given a departure mass from Earth orbit of 250 tons, the 'dry' mass that reaches Mars orbit is 100 tons.
Rated drive power is 15 megawatts. The propulsion system (thrusters and power supply combined) is allowed 50 tons, half the 'dry' mass of the ship. This corresponds to a power density of 0.29 kW/kg - or, putting it reciprocally, 3.45 kg of drive mass - thrusters and power supply - per kilowatt of drive power output. As a gearhead reference point that corresponds to 5.65 lbs per horsepower.
I allow 30 tons for fuel tankage, keel structure, and general equipment. Which leaves just 20 tons for the gross payload - life support hab, stores, and crew. This ship is half engine, not a very balanced design. But this is what you need if drive power density is limited and you want to get to Mars in a few months.
My math fu is not equal to the task of actually determining travel time to Mars for a given mission delta v. But this handy delta v calculator, can do it. According to the calculator, burns totaling 27.33 km/s of delta v are needed to get from a high (100,000 km) Earth orbit to low (500 km) Mars orbit in 90 days.
The calculator assumes a brisk 10 milligees of acceleration. The sims have a much more modest acceleration, averaging half a milligee. (And if the drive is solar electric its acceleration performance will be halved at Mars distance from the Sun. Therefore the calculator estimate quite optimistic, but clever mission design could probably squeeze out some improvements.
At least, to a first approximation, this provides some idea of what it takes to reach Mars in a few months.
But the mission profile is for a one-way trip, implying that the vehicle must refuel at Mars orbit. One day this may be a routine operation, but it will certainly not be routine the first time. Really we should be capable of a round trip with onboard propellant - requiring twice the mission delta v, about 55 km/s.
A second sim shows a lighter and more powerful drive engine, close to the 1 kW/kg figure of merit, putting out 30 megawatts and providing twice the specific impulse. This more powerful engine has a mass of 30 tons, allowing 40 tons gross payload. Realistically speaking this is close to the minimum performance requirement for a practical Mars craft - which is why 1 kW/kg is regarded as the figure of merit for fast space propulsion.
This spacecraft is strictly a crew transfer vehicle, intended to get the crew from high Earth orbit to low Mars orbit and back. Electric drive is completely unsuited to planetary landings, so any mission profile like Mars Direct is ruled out. Everything needed to land on Mars, live and work there for a time, and return to Mars orbit, can be sent on a slow orbit.
Since the mission departs from high Earth orbit, there must be a prior phase in which the ship, assembled in low orbit, spirals out, then is met by a crew ferry. In principle the ship could return to high Earth orbit and be met there. More likely at least on early missions, the payload will include an Earth return capsule for the crew (and samples of Mars material), with the interplanetary bus being expended.
This post is conceptually quite incomplete - really it only talks about the interplanetary bus. But I want to get it posted, so here it is.
The image of Mars' surface is from Astronomy Picture of the Day.