\section{Target System Summary}
 This chapter has presented conceptual designs of components to generate pions 
by bombarding a jet of mercury with high-energy protons, and then to capture 
the pions with a solenoidal field that bends the pion trajectories into helices 
that fit within the 0.15-m-diameter solenoid bore. 

The high-field region is 
0.6~m long, with a peak field of 20~T.   Downstream the field drops
adiabatically
by a factor of sixteen to 1.25~T over a distance of 18~m, while the bore 
increases by a factor of four. The mercury jet is 1~cm in diameter, with a speed of 
30~m/s and a tilt angle of 100~mrad relative to the axis of the magnetic field. 
An analytical estimate predicts that the jet should enter the target region 
with little deceleration and deflection.  However, these calculations suggest 
that the jet must not encounter any strong field gradient, if it is to avoid 
excessive shear and distortions in shape.  We allow the field to droop only 
$\approx 5\%$ over the 0.6-m-long target region.  Confirmation of the need for 
field uniformity comes from preliminary results from FronTier, a sophisticated 
hydrodynamic code that can track the free interface of the jet as it deforms 
in a magnetic field or breaks up from shock waves.

Finite element analysis predicts that pressure waves from the instantaneous 
heating of the 
mercury to several hundred degrees by the proton beam will splatter the jet 
completely. To replenish the 0.6-m-long jet in only 20~ms, the desired 
time interval between proton bunches, dictates the 30~m/s jet velocity.

Radiation emanating from the target is intense.  The computer code MARS predicts 
that the neutral flux rate through the beam pipe is up to $3 \times 10^{20}$
per cm$^2$ per year for neutrons, and an order of magnitude higher for 
gamma rays.  Charged particle flux rates are $10^{20}$ per cm$^2$ per year for hadrons and for electrons. The power dissipation is up to 2~W/g and the total 
radiation dose up to $4 \times 10^{10}$ Gy/yr.  These levels require shielding 
of many components, such as the pion capture magnet.

The pion capture magnet system is a hybrid, with many coaxial superconducting 
coils and a resistive insert.  %The system poses severe engineering challenges. 
The system stores 600~MJ, with a superconducting coil that generates 14~T in a 
bore of 1.3~m.  The resistive insert receives radiation so intense that only 
ceramic insulation will survive.  The baseline design for the insert uses 
water-cooled hollow conductor insulated with a layer of magnesium oxide 
between its copper conductor and sheath.  To generate 6~T in a large 
volume, the coil consumes 12~MW and requires many conductors in parallel in 
each layer to limit the hydraulic path length.  For a design lifetime of 
many years rather than a few months, the bore accommodates a layer of 
water-cooled tungsten carbide $\approx 10$~cm thick to attenuate the radiation 
by a factor of 30.  The pion capture magnet employs superconducting coils
of two types.  High-field, large-bore coils employ cable-in-conduit conductor. 
The lower-field, smaller-core coils that ramp the field down to 1.25~T  
employ Rutherford cable (as do the phase-rotation coils farther downstream). 
All coils require shielding to limit the power deposition to $< 1$ W/m$^3$, 
to avoid quenching, and to limit the radiation dose to $< 10$ MGy/yr
to enable organic insulation to survive.

Additional engineering challenges are mercury containment, mercury jet capture 
and diffusion, beryllium-window integrity and remote handling.  The computer 
code MCNPX predicts nearly 2~MCi of activation after only a hundred days of 
operation, with 105~Ci remaining after 30 days of cool-down.  The remote 
handling for maintenance and repair must deal with masses up to 45~tons and 
with components with limited accessibility.  All of these components will 
benefit from additional research and development.  Nevertheless, all aspects of the technology appear feasible.