This article presents some possible reusable launcher design options in the context of a scenario that might lead to a technological convergence between space vehicles and high-speed global aviation.

The ideal reusable launcher would be analogous to a normal aircraft in that it would be able to take off from many possible bases on Earth, enter the desired orbital plane, accelerate to orbital velocity, release its payload, de-orbit, dissipate its kinetic energy and land at the take-off base to be readied for its next flight after a quick turn-around. The vehicle would maintain physical integrity during a mission, would be flown 'very often', and its cost per flight would be compatible with the value of the missions undertaken. The vehicle would also be able to abort its mission any time in the case of mishap and land intact with its payload. This ideal vehicle is the 'single-stage-to-orbit reusable rocket launcher', or SSTO-RRL.

The value of the 'single stage' concept is that the vehicle maintains its integrity throughout its lifetime. This integrity, which results in the minimisation of the inherent interfaces, is a major factor in the vehicle's dependability. Indeed, the majority of today's practical transport systems are also 'single- stage' vehicles, which include aircraft, ships, cars, etc.

So far, however, all space launchers are staged to orbit and expendable (with the exception of the US Space Shuttle System) and there is no SSTO-RRL as yet, despite its apparent desirability. Clearly, something must be standing in the way of achieving this 'ideal' vehicle.

Because rocket propulsion is mandatory to accelerate to orbital speed in vacuum, the most logical design option is to use rocket propulsion from take-off until orbit insertion. Both gravity and drag losses must be overcome on the trajectory to orbit. The ideal velocity increment, Delta V, required from an SSTO-RRL is then about 9000 m/s in order to reach a Low Earth Orbit (LEO). All further considerations concentrate on reaching LEO, because this is the most difficult part of gaining access to space and the major hurdle to be mastered in terms of reusability.

The mass that can be accelerated into orbit using rocket propulsion is given by the equation: M 1 /M 0 =exp ( -Delta V /V E ) where M 0 is the mass at take-off, M 1 is the mass which has received the ideal velocity increment Delta V, and V E is the ejection velocity of the rocket engine.

For a given Delta V, which is mission-imposed, the mass ratio M 1 /M 0 increases with increasing V E . The highest practical rocket ejection velocities are achieved by burning hydrogen with oxygen in a combustion chamber and ejecting the produced gases through a convergent/divergent nozzle. When averaged over the trajectory, the exhaust velocity V E is in the order of 4000 m/s. The corresponding mass ratio to reach LEO is: M 1 /M 0 =exp(Delta V /V E ) = exp ( 9000/4000) = 0.1054 = 10.54%

This means that 100 10.54 = 89.46% of the take-off mass must be made up of propellants, and that just 10.54% of the trade-off mass remains available for the tankage to contain the propellants, the engines, the structures, the equipment, and last but not least the payload. Assuming that the payload is 1% of the take-off mass, the mass fraction left available to build the complete launcher is 10.54 1 = 9.54%. This is not very much, and we have already assumed the highest performing propulsion system possible (cryogenic propulsion).

Mass ratios of the order of 10% are achieved today with expendable rocket stages, but are a nearly impossible requirement for a reusable vehicle, which is more stressed during its lifetime and which must carry the additional provisions for recovery and reuse. Because today's rocket propulsion is already quite near to its theoretical maximum, the whole burden of achieving such small mass ratios falls on the materials from which the launcher is built. Such launchers are, however, still studied within the FESTIP initiative (Fig. 1) in order to define the areas in which technology advances would be most effective.

If the mass ratio allocated to the SSTO-RRL could be larger, the vehicle might become more feasible or more robust to use. The options offered by the rocket equation for increasing the mass ratio are to increase V E ,or to decrease Delta V, or a combination of the two.

Figure 1. A possible single-stage-to-orbit launcher studied
within ESA's FESTIP initiative

If V E were increased by 10% from 4000 m/s to 4400 m/s, the mass ratio would become

M 11 /M 10 = exp ( 9000 / 4400) = 0.1293

Again assuming 1% for the payload, the mass available for the vehicle becomes

12.93% 1% = 11.93%

which represents a percentage increase of

(11.93 9.54) / 9.54 = 25.1%

with respect to the nominal SSTO-RRL, which is quite a significant gain.

Increasing V E is equivalent to finding a propulsion system which produces the required thrust while consuming less propellant. For example, the apparent V E of propellants carried aboard a rocket increases when some of the products ejected are taken from the ambient air instead of having to be carried onboard from the start. In general, however, increasing the V E of a rocket motor with the help of the atmosphere is a complex problem to which Europe cannot expect an early solution. A typical configuration of this type, shown in Figure 2, is the SKYLON (UK), derived from HOTOL.

Figure 2. A possible single-stage-to-orbit launcher with combined
air-breathing and rocket propulsion

Turning now to the effect of a reduction in velocity increment, if the required Delta V were reduced by 10% from 9000 to 8100 m/s, the mass ratio would become

M 1 /M 0 = exp ( 8100 / 4000) = 0.1320

If we still want a 1% payload in orbit, the payload MP0, ejected at 9000 8100= 900 m/s short of orbital velocity, must carry an additional propulsion system. Assuming a storable bipropellant system having an effective exhaust velocity of V E2 = 3000 m/s, the mass fraction for the payload is

MP 1 /MP 0 = exp ( 900 / 3000) = 0.7408

The mass ejected 900 m/s short of orbital velocity must therefore be 1 / 0.7408=1.350% of the launcher take-off mass to again have a final payload mass in orbit of 1%. The net mass ratio remaining for the launcher itself becomes

13.20% 1.350% = 11.85%

which represents a gain of

(11.85 9.54) / 9.54 = 24.2%

with respect to that required for the pure SSTO-RRL.

Therefore, reducing Delta V is as effective in creating a mass margin as is increasing V E. In more general terms, reducing Delta V leads to having more than one stage to reach orbit. As each stage has to deliver only a fraction of the total Delta V, its mass fraction can increase and a more robust design becomes possible.

** (a) The SSTO-AL(air-
launched)-RRL*** Principle:* A carrier aircraft
takes the SSTO-RRL to a certain altitude and speed, thereby
reducing the remaining DV required to reach orbit. The Interim
HOTOL carried on-top of an Antonov-225 aircraft is one such
example (Fig. 3).

Figure 3. The Interim HOTOL launched from the Antonov 225
aircraft

*Drawbacks*: The carrier aircraft has a limited payload
capability, which constrains the SSTO-RRL's mass at separation.
As a result, an air-launched SSTO-RRL becomes as difficult to
build as a ground-launched version for which no take-off mass
limitations apply. This comment is applicable to any upper stage
of any Two Stage To Orbit (TSTO) system, because large vehicles
are more effective and have a lower overall mass-ratio
capability. Air-launching also involves a dangerous separation
in the presence of aerodynamic flows.

*Advantages*: Air-launching gives launch azimuth
flexibility and provides a self-contained flying launch base.

*Conclusion*: Not a promising solution for Europe.

**(b) The TSTO-RTL (return to launch site) and DRL(down-
0range-landing)-RRL**

* Principle*: Here, a
reusable rocket first stage carries a reusable (or expendable)
rocket second stage. In the RTL mode, the first stage returns to
the launch site after separation. RTL is practical as long as the
first stage does not re-enter too far down range, which
corresponds to up to 1/4 of the total Delta V required to reach
LEO. The FLS as studied by Aerospatiale for ESA in the early
1980s (Fig. 4), the RRL studied in the early 1990s (Fig. 5) and
the FSS 9 as studied within FESTIP (Fig. 6) are vehicle's of this
type.

Figure 4. The semi-reusable future launching system studied by
Aerospatiale in the early 1980s

Figure 5. A semi-reusable rocket launcher as studied by
Aerospatiale in the 1990s

Figure 6. A fully reusable two-stage-to-orbit rocket launcher as
studied within FESTIP

In the DRL mode, the first stage provides a larger share of the total Delta V and is allowed to land down-range of the launch site after separation and re-entry. The first stage returns to the launch site at a later time. The TARANIS, studied by Aerospatiale for ESA in the early 1990s (Fig. 7) was this type of vehicle.

Figure 7. TAKANIS: a fully reusable rocket launcher with down-
range landing of the first stage

*Drawbacks*: There are now three aerodynamic
configurations to be controlled - the first stage, the second
stage when reusable, and the composite - two vehicles to be
developed and qualified, and the amount of operations is
increased. Staging is optimised for a given technology level and
prevents one from deriving full benefit from any subsequent
technology improvements. Because of the mass limitations on the
second stage, its reusability represents a penalty rather than
a cost advantage. A more effective solution is that of an
expendable second stage, which could be considered an interesting
interim solution.

When implemented with today's technology, the TSTO-RTL/DRL-RRL is not much cheaper in use than an expendable launcher, but it can already offer higher reliability and safety. Launch/in-flight abortability remains limited.

*Advantages*: Feasible in the near term, if needed.

*Conclusion*: Not an interesting solution for Europe, in
that the expendable Ariane-5 is already a modern launcher with
significant evolution potential.

**(c) The SOSS (Sub-Orbital Single Stage) OA (Once-
Around) RRL**

* Principle*: The down-range
capability of the first stage is now extended to the point where
it can complete one trip around the Earth so as to land back at
the launch base (Fig. 8). The first stage therefore provides
almost the total Delta V needed to reach orbit and the payload,
ejected in vacuum but still at sub-orbital speed, achieves
orbital velocity with its own propulsion system. This principle
is comparable to that of the SSTO-AL-RRL, except that now the
Delta V complement provided by the carrier aircraft (a large and
expensive item) is provided by the payload itself (a small
system) and the large sub-orbital stage itself retains the
freedom of unconstrained mass at take-off.

Figure 8. The trajectory of a once-around Earth sub-orbital
single-stage reusable launcher

*Drawbacks*: The payload must provide its propulsion into
orbit (from 300 to 3000 m/s) to compensate for the Delta V
deficit of the sub-orbital stage. Not all payloads might welcome
this, but this requirement could be taken into account when
designing them.

*Advantages*: For the payload: For the majority of
payloads, LEO is not the ultimate destination and many of them
already carry an integrated propulsion stage to acquire higher
energy orbits. Adding the need to provide the initial impulse to
reach orbit therefore does not represent an excessive penalty.

*For the launcher*: The launcher is now back at the
launch site after one revolution. The Delta V needed to
circularise and to de-orbit is saved, there is no launcher in-
orbit phase with the associated functions and mass, and a high
launcher efficiency is therefore possible.

*Conclusion*: The author believes that the SOSS-OA-RRL
(sub-orbital single-stage once-around reusable rocket launcher)
can be a very promising solution, the particular features of
which are discussed below.

Let us start by defining a large reusable stage with a given take-off mass, of say 500 tons. Since we do not yet have the technology to achieve the mass ratio necessary for true SSTO operation, we must design the launcher to provide the largest Delta V it can, and leave it to the payload to provide the complement needed to achieve orbital conditions. Keeping the take-off mass constant allows us to freeze the launcher's external shape, so that the aero-dynamic database remains valid once established, and the installed thrust can be fixed. The launcher tankage is designed for the full volume of propellants ultimately required for SSTO, and the payload bay is given the volume needed by the largest possible combination of payload with its upper-stage propulsion.

We can now trade-off propellant masses between the launcher and the payload whilst keeping the total take-off mass constant. Depending on the sharing of Delta V between the launcher and the payload, we can achieve a range of payload-in-orbit capability. This approach therefore combines the flexibility and performance robustness of a pure TSTO system with the operational benefits of a single-stage vehicle. We refer to this approach as 'internal staging'.

The SOSS-OA-RRL, when already dimensioned for ultimate SSTO capability, can therefore become an interim single-stage vehicle with which practical experience can be accumulated early in a revenue-generating utilisation environment, whilst still preserving the possibility to upgrade it as technology progresses, without affecting its overall layout. With improving technology, the launcher dry-mass or the propellant mass needed will decrease, the launcher will reach higher speeds, and the propulsion requirements on the payload will decrease accordingly, until a pure SSTO capability becomes possible, should this indeed prove to be an economically viable option.

In a first approximation, the cost of using an SOSS-RRL is constant and independent of the mass of the payload with propulsion it carries, because its take-off mass is constant. The same SOSS-OA-RRL is used for each payload, irrespective of its mass, the only adaptation being the correct loading of propellants and the calculation of the corresponding trajectory. This repeated utilisation should lead to significant cost savings as experience accrues.

Our SOSS-OA-RRL now resembles a truck which has a maximum payload, but can also operate profitably with a partial load.

The velocity deficit of the SOSS-OA-RRL with respect to orbital velocity cannot be too large if we want the reusable launcher stage to complete one turn of the Earth and return to the launch base in a non-propulsive manner. The permissible deficit will depend on the vehicle's aerodynamic qualities during the high- speed re-entry phase: a greater lift-to-drag ratio at hypersonic speed permits a greater passive down-range capability, which permits a larger allowable velocity deficit with respect to orbital speed, which in turn permits a greater launcher dry-mass (at the expense of increased aerothermal problems). Suitable shapes for the hypersonic glide-back flight are lifting bodies (Fig. 9) or waveriders (Fig. 10).

Figure 9. A lifting-body sub-orbital, once-around reusable rocket
launcher studied within FESTIP

Figure 10. A generic 'wave-rider' shape as studied within FESTIP

An SOSS-OA-RRL which needs a down-range capability to return to its base therefore makes constructive use of the kinetic energy it received during ascent, instead of simply destroying it as is the case for re-entry from orbital speed.

If the velocity deficit is such that even good aerodynamics do not allow a complete once-around glide, the stage can land after a partial orbit, but then there is the problem of returning it to the launch base, or we can equip it with an auxiliary propulsion system for a powered fly-back to complete the once- around. This powered fly-back could either be ensured by a scramjet for high-altitude, high-speed flight (although there might be thermal problems), or as a subsonic cruise on turbojet power (aerial refuelling with kerosene is a possibility). Again, the better the aerodynamics, the less penalising the cruise-back becomes.

Now that we have a reason to require good aerodynamics for the return flight of the SOSS-OA-RRL, the next step is to use the installed aerodynamics for horizontal take-off (HTO) and lifting ascent on rocket power only. Indeed, the integrated velocity losses with HTO are a little less than with vertical take-off (VTO) on rocket power. A conventional SSTO-RRL must, however, rely on VTO because its required low mass fraction does not permit the implementation of lifting surfaces for ascent.

For a reusable vehicle, however, VTO has unpleasant implications. With VTO, the installed thrust which translates into number, mass and cost of engines must lift the vehicle off the pad, must provide the thrust needed to accelerate it, and there must also be a thrust margin for some engine-out capability. The total installed thrust then corresponds to about 1.5 times the take-off weight of the vehicle, only about a third of which does useful work in accelerating the vehicle during the initial phase of the mission. This installed thrust becomes excessive towards the end of the ascent and engines have either to be throttled back (cost) or shut-off (dead mass into orbit).

Because rocket engines must be located at the base of the vehicle, a high installed thrust shifts the centre of mass of the empty vehicle to the rear and compromises must be made in the aerodynamic shape in order to maintain a controllable vehicle at re-entry.

Even when planning for engine-out situations, VTO remains a dangerous phase and while this is normal for an expendable launcher (if it has to fail, it does not matter where it fails), it is far from ideal for a reusable vehicle. Indeed, until there is enough dynamic pressure to provide aerodynamic control authority, the reusable launcher behaves like an expendable vehicle and the range-safety and vehicle destruct may have to be retained.

Finally, VTO requires a heavy ground infrastructure which is both costly and limits the number of launch bases that one can build.

The SOSS-OA-RRL launcher, which needs aerodynamics for its return flight (and which can afford it because of its higher permissible mass fraction) allows one to exploit the Delta V savings of HTO. However, HTO also has more fundamental advantages. With HTO followed by lifting ascent, the weight of the vehicle is carried by the lift due to the forward speed and only the drag generated by that lift has to be compensated for by rocket thrust. Depending on the aerodynamic shape, the drag generated by the lift is only one half to one quarter of the vehicle's weight. Rocket thrust to accelerate the vehicle at the rate of 0.5 g at take-off is sufficient, so that the total rocket thrust needed by an HTO vehicle is about 0.5 + 1/2 to 1/4 =1.0 to 0.75 times the weight of the vehicle at take-off. This reduced thrust budget leads to associated reductions in engine mass (which improves the aerodynamic centering) and cost.

HTO also provides an increased engine-out capability and therefore greater safety, at least when implemented as follows. The vehicle can rest on a passive trolley, on a linear ground track a few kilometres long. The vehicle accelerates under its own rocket power until the appropriate velocity is reached, at which point its nose is pushed upwards mechanically causing the vehicle to lift off from the trolley and continue its ascent. Operational advantages of this approach are:

- until the vehicle leaves its trolley, the take-off can be aborted
- once minimum control speed plus a predetermined margin is reached, the take-off can be initiated, but should there then be a loss of thrust the excess velocity is used to stabilise the vehicle and the only thrust needed is that to compensate for the drag, which is about 1/2 to 1/4 of the total installed thrust; the vehicle can dump its oxygen to reduce its weight, thereby increasing the thrust margin, allowing it to turn back and land safely.

With this scenario, rocket-powered HTO could offer the same thrust and safety margins as are now customary for multi-engined aviation.

Finally, the HTO lifting ascent phase prolongs the time spent in the Earth's atmosphere by several minutes, which might allow some form of air-breathing rocket propulsion to be added to the vehicle at a later date, thereby improving the apparent specific impulse of the propulsion system and further increasing the allowable mass ratio.

By now, our proposed SOSS-OA-RRL has quite striking parallels with conventional aviation: horizontal processing, horizontal take-off, horizontal ascent, gliding or powered hyper-sonic flight, horizontal landing, and similar engine-out tolerances. The vehicle can cover all speed ranges from low subsonic to nearly orbital, and can evolve through the whole depth of the atmosphere. But such a vehicle might have other applications too, as it is not obliged to make a full revolution of the Earth: it could then serve as a high-speed global transport vehicle.

Indeed, high-speed commercial transport (HSCT) programmes in the USA, Japan and Europe are already devoting significant funds to solving the fundamental technology problems, and the approach being pursued so far envisages cruising at higher speeds (up to Mach 5.5), as a direct extrapolation of today's lower speed aviation. At such speeds, however, global range trip times remain long, the aircraft cruises in contamination-sensitive layers of the upper atmosphere where its engines deposit emissions, the aircraft generates a sonic-boom carpet on the ground, which makes supersonic overflight of land masses unlikely, the aircraft faces severe thermal soaking problems, the propulsion still has to be developed to civilian standards, etc.

The boost-glide approach of the SOSS-OA-RRL is a logical alternative for such global travel: the vehicle is accelerated to high speed (Mach 10?,15?) by rocket (not necessarily O 2 /H 2 ) or by rocket plus air-breathing propulsion, follows a ballistic arc in near-vacuum, re-enters and dissipates its kinetic energy during an equilibrium glide at hypersonic speed and at high altitude. Once this energy has been dissipated, the vehicle can make a conventional landing like a normal subsonic aircraft.

The advantages of boost-glide for global travel are:

- The boost phase is a pure acceleration phase with little energy loss, and takes place over a small area only. Rocket exhausts contain no nitrogen oxides as no air is combusted.
- The ballistic phase is outside the Earth's sensitive atmosphere and therefore has no detrimental effect on the planet.
- The hypersonic glide is at such high altitudes that the sonic boom does not reach the ground and therefore land masses could be overflown at hypersonic speed.
- The vehicle's high speed allows it to benefit from centrifugal force in order to reduce the required lift, which in turn reduces drag and provides a greater operating range for the same initial energy.
- Global travel times are much reduced compared with the supersonic-cruise approach.

Assembling the features described, one can conclude that, with today's emerging technologies, launcher reusability offers the opportunity to serve both the space-launch and global-travel markets with one type of vehicle: the sub- orbital single-stage reusable rocket launcher. The vehicles themselves would differ in detail, depending on their exact roles, but they would rely on a common technological base.

Such synergy would be to the benefit of both communities: it would enlarge the technology base available for space activities, it would spread the resulting development costs, it would amortise more rapidly all investments made, and it would stimulate a new approach to travelling around our planet, which in turn could prove a strong motivator for the younger generation to pursue a scientific/technical career.

Space has the potential to become a very large industrial park, stimulated by more routine and less costly access to space in the same way that global markets have been facilitated by the growth and falling cost of today's Earth-bound transport systems. Assessment of the possibilities for making this happen deserves commensurate spending for exploratory work. ESA's FESTIP programme will hopefully contribute to this goal.