\section{Introduction}
In this section we summarize the key R\&D activities required to validate
the design concepts described in this Neutrino Factory Feasibility Study.
Topics will be covered in the order in which they appeared in the facility
descriptions given earlier in this document. Items covered here fall into
two categories: i) those required to validate or improve the components that
drive the fabrication costs of the facility, and ii) those required to
address the performance and/or feasibility of fabrication of particular
components. In the first case, R\&D will mainly involve hardware fabrication
and testing without beam. In the second case, performance tests with beam
may be required in addition to prototyping. For each hardware area, the main
R\&D topics will be listed in the context of the two categories above. 

The R\&D items listed here fall into the broader R\&D effort of the Neutrino
Factory and Muon Collider Collaboration (MC). A five-year R\&D plan,
currently under way, has been completed by the MC and is available~\cite{RandD:ref1}. We will not repeat that
information here. What we cover below are topics that have arisen in the
context of this---and, in some cases, the previous---Feasibility Study.

It is important to note that much of the hardware development effort
envisioned here requires different people at different institutions. Thus,
there is no fundamental reason that the program cannot proceed in parallel
on several fronts. Indeed, it {\it must} proceed this way if we are to
complete the R\&D tasks in a reasonable time frame. Clearly, however, our
progress requires funding commensurate with the
program needs; this is the resource over which we have the least control.

\section{Proton Driver}

The upgrade of the AGS to reach higher intensity is relatively minor, as it
has already operated at 70\% of the design intensity specified here. The
aspect where additional work is required is related to the need for short, 3
ns, bunches. The peak current required in this case, about 400 A, is seven
times higher than what the AGS has achieved to date. Efforts to reduce the
ring broadband impedance below 10~$\Omega $ will be worthwhile. For example,
development of bellows shields capable of operating reliably in this beam
current regime should be examined. The new B Factories (PEP- II and KEKB)
are operating in similar regimes of peak and average beam current without
experiencing reliability problems with bellows. In parallel with this
effort, it will be important to explore other means to mitigate transverse
instabilities in the AGS, {\it e.g.}, by introducing tune spread by means of
octupoles.

For the new rf cavities, the 4L2 ferrite material must be characterized in
the appropriate frequency regime. (The use of 4M2 ferrite, which has been
used elsewhere, is an acceptable fallback solution if the ''standard'' AGS
cavity ferrite does not have the required properties.) 

The trade-offs between using a thinner vacuum chamber wall and stronger
sextupoles should be studied. In addition, it will be necessary to examine
all of the power supply tracking algorithms to make sure they will work
properly at a three-times-higher ramp rate. All of these kinds of issues
have been solved previously in other accelerators, so there are no real
unknowns here.

Experimentally, the AGS should continue with efforts to produce short
bunches, to understand the present limitations and make sure that the ring
impedance is well understood.

\section{Target System}

The main technical issue to deal with here concerns the survivability of a
mercury jet in a 1 MW proton beam. Experiments are already under way at the
AGS to study this, and an effort to predict the behavior of the mercury jet
via simulations is proceeding in parallel. Ultimately, yield measurements to
validate the MARS predictions should be carried out. These will include a
(pulsed) 20-T solenoidal field.

Testing of a mercury jet in a high magnetic field is already in progress in
Europe
%~\cite{RandD:ref2} 
using a 13-T magnet at Grenoble. If
necessary, tests could be repeated at the full 20-T field at the NHMFL.

The potential of failure fatigue for the jet nozzle must be studied. There
is a pressure shock traveling back toward the nozzle after each proton beam
pulse. A verification experiment for the shock effects, along with
corresponding simulation work, should be carried out to permit a realistic
means to mitigate it. Means to ``pulse'' the mercury jet should also be
examined as a possible workaround.

Alternative designs, such as a ``band'' target (see Section~\ref{APP-band}), should be examined. In
addition to the mechanical issues of the band itself, compatibility with the
solenoid configuration envisioned for the target must be assessed via a
solid engineering concept.

Studies of the cost-benefit tradeoffs between capture efficiency and
magnetic field should be made to optimize the target solenoid field. Present
indications are that the penalty of a decrease from 20 T to 18 T is minor in
terms of intensity, but the corresponding studies of magnet cost must be
carried out. Studies of more optimal conductor for the hollow-conductor
magnet are also needed. A conductor having wrapped ceramic insulation should
have a more favorable power consumption than the MgO insulated conductor,
while providing the same magnetic field. Studies of the alternative
Bitter-magnet technology are also needed, focusing on issues of lifetime.
The Bitter magnet is more efficient than a hollow-conductor magnet, but its
resistance to corrosion in the high radiation environment must be studied.

There are a number of technical issues to consider in completing the design
of the target containment area. These include:

\begin{itemize}
\item  {\it Beam stop design}. A study of various materials should be
undertaken to compare the advantages of low-$Z$ and high-$Z$ materials.
Issues include secondary particle showers, residual activation, and decay
heating.

\item  {\it Beam containment window design}. Adequate cooling designs must
be developed, possibly including a combination of bulk coolant flow and edge
cooling. If water cooling is required, adequate machine-protection
interlocks must be developed. At present, beryllium looks like the most
promising material. There is some commonality between the windows for the rf
cavities and the beam containment that should be exploited. Since these are
replaceable components, techniques for remote replacement must be developed.

\item  {\it Component cooling}. Activation of coolants, primarily light
water, must be studied. In particular, $^{7}$Be and tritium must be
considered in the design of the cooling system.

\item  {\it Radiation damage}. The effect of radiation damage on the iron
magnet ``plug,'' the hollow-conductor copper coils, and the superconducting
coil must be studied. Effects of intense radiation on mechanical properties
(strength, elasticity) and buildup of corrosion products must be assessed.
Materials tests in this context are already in the planning stages.
\end{itemize}

There are several R\&D issues related to the facility configuration and
design of the nuclear shielding for the solenoids. The first facility issue
deals with the location of, and access to, the target and magnet support
systems, namely, vacuum pumps, ducts, valves, cryogenic lines, electrical
cables, and diagnostic equipment, to ensure that they are readily
maintainable. A target hot cell is already configured with access and
maintenance in mind; a more detailed iteration of the facility design would
accomplish the same for the other support systems. The second facility issue
is the extent to which remote handling capability and equipment are needed
downstream from the target/capture region. An extrapolation of the shield
analysis that was done for the floor shield over the tunnel, $-0.8<z<36$~m,
indicates that similar requirements apply downstream. If verified, this
requirement would have an impact on the overall facility design and cost.

There are several issues that deal with the nuclear shield design. These can
be addressed with R\&D activities that simultaneously address mechanical and
thermal questions. It has been determined that the optimum shield for the
high-field solenoids is 80\% tungsten-carbide, 20\% water, so the shield
design is based on using tungsten-carbide balls. Scale model tests are
needed to investigate how to distribute the balls in a homogeneous matrix,
and to assess properties such as pressure drop and heat transfer
coefficient. This shield is a costly component, and it is important for it
to be efficiently designed.

\section{Phase Rotation and Capture}
\label{APP-mini-cool}
In this Study, the induction linac (IL) design is closer to present day
experience than was the case for Study-I, so the technical uncertainty is
lower. Nonetheless, the gradients required are high, so a prototype
induction cell, along with its magnetic pulse compression system, should be
fabricated. Furthermore, there are several possibilities that might lead to
a more cost-effective implementation. Development of less lossy (thinner)
amorphous alloys should be undertaken, in conjunction with industry. Being
able to use a mass-produced material with acceptable loss properties will
result in lower capital costs initially, and lower power costs for the
operating facility as well. Candidate materials need to be tested to
validate the properties on which the design is based. In addition, the
branched magnetics concept should be developed. If it is acceptable to drive
a single core with two independent unipolar pulsers, it would eliminate one
induction linac in our design. Radiation tests on the Mylar core insulation
should be made, to be sure there is no degradation over the expected life of
the facility.

The design of the IL core is strongly influenced by the internal
superconducting solenoids, in the sense that the inner diameter of the core
is set by the need to avoid the fringe field from the solenoid. Quantifying
the effects on the core of the solenoid fringe field must be done. In
addition, means to reduce the solenoid fringe field, thereby permitting the
IL core inner diameter to be reduced, should be examined as part of a
cost-benefit tradeoff study.

For the capture area, studies of the radiation heat load need to be refined,
and extended through the IL region. The shielding requirements, especially
for the upstream solenoids, have a strong impact on the inner diameter, and
hence the cost, of these magnets, and an optimization is required. It would
clearly be prudent to consider the future upgrade to 4 MW in this
regard, as it would be undesirable to have to upgrade the magnets and
shielding later.

The present mini-cooling absorber design is not optimized. The main
requirements for mini-cooling are: {\it i})~energy loss equivalent to that
of $2\times 1.75$\thinspace m of liquid hydrogen; and {\it ii})~low multiple
scattering. Liquid-hydrogen mini-cooling, while straightforward, is
technically, complicated---undesirably so. Simpler solutions involve
non-cryogenic liquids or low-$Z$ solids. It is clear that some R\&D is
called for to flesh out these options. Table~\ref{tab:minicooling-appendix}
summarizes the lengths and corresponding radiation-length fractions for
liquid hydrogen and various alternative materials. 

While hydrogen minimizes scattering effects, it is likely that solid lithium
or beryllium would also be acceptable. (Lithium hydride presents practical
difficulties since it is neither commercially available nor readily
manufactured in large, shaped pieces.) Simulations show that, compared with
liquid hydrogen, solid lithium mini-cooling absorbers reduce the number of
muons per proton by only about 5\%, and beryllium causes only about a 10\%
reduction. These performance degradations are small, and they can probably
be avoided by raising the solenoidal field somewhat to compensate for the
increased scattering. (This would entail reoptimizing the front end.)
Stronger focusing will increase the cost of the solenoids, but will reduce
the size of the beam, allowing smaller absorber diameter. An overall
optimization of the system, involving both simulations and engineering,
should be done.

\begin{table}[tbh]
\caption{Comparison of possible minicooling absorber materials.}
\label{tab:minicooling-appendix}
\begin{center}
\begin{tabular}{|lcc|}
\hline
Material & Length & Radiation length \\ 
& (cm) & (\%) \\ \hline
LH$_2$ & 175 & 20 \\ 
LiH & 38 & 35 \\ 
Li & 57 & 37 \\ 
CH$_4$ & 49 & 45 \\ 
Be & 17 & 48 \\ 
H$_2$O & 25 & 70 \\ \hline
\end{tabular}
\end{center}
\end{table}

The roughly 5~kW of power dissipated in each mini-cooling absorber
appears manageable. Cooling tubes affixed to the large ($\sim $1\thinspace m$%
^{2}$) perimeter surface can easily transfer such heat. Conductive heat
transfer through the material, from the core to the periphery, requires only 
$\sim 10^{\circ}$C temperature rise, small compared with the
melting points of lithium and beryllium ($186^{\circ }$C and 
$1350^{\circ}$C, respectively). Water, freon, or some other
convenient refrigerant might be suitable, with a choice other than water
preferred in the lithium case to reduce the risk of reaction should cracks
develop in the cladding.

On the practical side, the feasibility and cost of fabricating large
cylinders of these materials must be evaluated. Preliminary contacts with
manufacturers~\cite{Diesburg-FMC},~\cite{Brush-Wellman} suggest that these are not
fundamental problems. After design work, fabrication of a prototype disk,
followed by bench (and perhaps beam) tests of its thermal performance, would
be desirable.

\section{Buncher and Cooling}
\label{APP-NCRF:grid}
The solenoid designs need to be cost optimized and the results put back into
the simulations. In particular, the forces on the focusing coils in Lattice
2, with its 1.65-m cell length, are quite high. Lowering these forces will
reduce costs. A somewhat longer cell length should help here.

Absorber development R\&D is well under way. \ Techniques to produce very
thin windows have already been developed. \ Pressure tests are planned to
validate the safety aspects of the design. \ Because the power density is
high, cooling of the absorber to avoid density fluctuations is challenging.
A program of fluid dynamics modeling and bench tests is under way, to be
followed by beam tests with 400 MeV protons at Fermilab.

Development of diagnostics is an ongoing process. Prototype devices of the
types mentioned in Chapter~\ref{BandPR:chap} must be built and tested. Some of these tests
will be carried out in conjunction with the rf cavity tests in Lab G at
Fermilab. Possible backgrounds from the cavity can be assessed this way.
Where possible, diagnostics devices will be tested in a beam, either the 400
MeV proton beam at Fermilab or possibly a muon beam at BNL or elsewhere.

Emittance exchange offers the potential of doubling the intensity of the
facility. This is a difficult problem, presently the subject of simulation
effort. This effort will be continued to see if an acceptable scheme can be
developed. \ If a good solution is found, hardware development will follow,
including new components, such as wedge absorbers, that are called for in
the design concept.


In practice, the most critical technical component of the cooling channel is 
the rf system. The rf peak power requirement is very high for the cooling 
channel; means to reduce this will yield large benefits. The main issue is 
to optimize the cavity design for minimum power requirements at the required 
gradient, and then optimize the cooling lattice design with a suitable cell 
length. (To date, we have always done this process in inverted order, 
leading to a non-optimal rf cavity design.) 
The normal conducting rf structures for the Neutrino Factory buncher and
cooling sections are challenging due to the high gradients required and the
large transverse dimensions of the incoming muon beam. The solution we are
pursuing is to close the beam iris with a conducting barrier of low-$Z$
material to restore the shunt impedance. Simulations indicate that thin
beryllium foils, or arrays of thin-walled aluminum tubes, restore the shunt
impedance while maintaining acceptable beam scattering. For a continuous
foil, the minimum thickness is determined by the power dissipation on the
surface. In vacuum, at close to room temperature, the heat can only be
removed by radial conduction through the foil to a water-cooled flange. This
produces a temperature gradient in the foil, with the maximum temperature in
the center. The result is a tendency for the center material to expand and
the foil to bow, detuning the cavity. This tendency can be eliminated, up to
a point, by arranging for the foil to be pre-stressed in tension. This keeps
the foil flat up to that temperature at which the thermal expansion exceeds
the pre-stress. Alternatively, the foils could be pre-bowed (to predetermine
the direction of motion), and the movement accommodated by tuning the
cavity. Such a pre-stressed foil has been simulated in ANSYS and
investigated experimentally in a series of tests on small foils at 805 MHz.
These will continue as part of the R\&D effort, including high-power testing
of a cavity with foils in the Lab G facility at Fermilab.

Other structures under consideration include grids of thin-walled tubes and
other fabricated structures, see Figs.~\ref{APP-RF:fg1} and ~\ref{APP-RF:fg2}%
. An advantage of closed tubes would be the ability to flow cooling gas
through the structure, potentially allowing larger apertures or less
material to be used. Simulations suggest that the grids provide adequate
isolation between cavities with tolerable scattering of the muon beam. The
tubes themselves cause local concentrations of the electric and magnetic
fields near their surfaces, but the kicks to the beam from this source are
estimated to be small compared with other transverse deflections. R\&D is
needed to develop all of these candidate structures and test prototypes
under realistic conditions. Manufacturing of pre-stressed foils large enough
for the 201.25 MHz cavities needs to be investigated further. Fabrication
technology for the arrays of thin-walled tubes also needs to be explored.

Cost-effective manufacturing methods must be developed for the 201.25 MHz
cavities themselves. We are contemplating processes such as spinning or cold
forming for the large cavity shells, and electron beam or laser welding for
the joining processes. Suitable windows, tuners and ancillary equipment must
be developed for the high-power and high-gradient regime we require.

Given the high rf power requirements, and the inapplicability of
superconducting rf due to the high magnetic field, it is interesting to
consider running the conventional copper cavities at lower temperature to
improve their conductivity and thus reduce the wall losses. Anecdotal
evidence suggests that wall losses may decrease by a factor of two at
liquid-nitrogen temperature, although hard data for actual operating
structures has not been forthcoming thus far. This would reduce the peak rf
power requirements, at the expense of increased refrigeration capacity. The
cost tradeoff between these two expensive systems will be evaluated. Up to
this point, we have taken care to maintain the possibility of
low-temperature operation in the design; none of the proposed hardware
configurations preclude this option.

\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA1.eps}
\end{center}
\caption{Grid of thin-walled tubes.}
\label{APP-RF:fg1}
\end{figure}
\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA2.eps}
\end{center}
\caption{Continuous array of tubes.}
\label{APP-RF:fg2}
\end{figure}

Large scale integration of the rf structures into the lattice will also be
the subject of ongoing R\&D. The close proximity of the rf cavities to
superconducting solenoids, the liquid-hydrogen absorbers, and the
instrumentation makes for some technical challenges and results in many
tradeoffs. For example, the diameter---and therefore the cost---of the
largest solenoid coil could be reduced by reshaping the center RF cavities,
but at the penalty of reduced shunt impedance. The shunt impedance is also
strongly dependent on the amount of longitudinal space available. Figure~\ref{APP-RF:fg3} shows how the shunt impedance per cavity, and per meter, varies
with length. We will continue to explore the cost minima of these tradeoffs.

The cost of rf power at this high level has prompted us to adopt a
multi-beam klystron (MBK), as our baseline power source for these studies.
MBKs have been developed at other frequencies for applications such as the
TESLA test program, and have been successful at meeting expected power
outputs and efficiencies~\cite{mike:ref1}, Figs.~\ref{APP-RF:fg5},~\ref{APP-RF:fg6},~\ref{APP-RF:fg7},~\ref{APP-RF:fg8}. Preliminary contacts with
tube manufacturers suggest that development of a 201.25 MHz MBK would be
technically feasible and economically viable, given the scale of the
Neutrino Factory. This type of source will be investigated further as part
of the ongoing R\&D plans. Other potential sources might include improved
tetrodes or {\it diacrodes} and other beam-based devices, such as inductive
output tubes (IOTs) or hollow beam tubes ({\it hobetrons}). Figure~\ref{APP-RF:fg9} shows a prototype high average power IOT~\cite{mike:ref2} .
Table~\ref{APP-RF:tb1} compares this to an equivalent conventional klystron.
We will continue to study these alternatives and watch developments in the
field. The cost and performance of power supplies and modulators have
improved in recent years due to developments in solid-state switching
devices (such as IGBTs and SCRs), and thanks to the intensive R\&D
activities for linear accelerators. We will continue to refine our proposed
design to take advantage of any further advances in this field.

The objectives of the R\&D plan for the Buncher and Cooling Channel rf
system are as follows:

\begin{itemize}
\item  Perform high power tests of the open- and closed-cell cavities in Lab
G at Fermilab

\item  Demonstrate that the required gradient can be achieved in the high
magnetic field

\item  Investigate the conditioning and performance of a cavity containing
beryllium foils, with varying levels of magnetic fields

\item  Investigate the necessity and effectiveness of anti-multipactor
coatings, such as TiN

\item  Study the effectiveness of foils, grids, and other assemblies
suitable for the 201.25 MHz cavity

\item  Investigate manufacturing methods for the 201.25 MHz cavity itself,
and for foils or other structures suitable for the large diameter iris

\item  Prepare a conceptual design for a high-power 201.25 MHz test cavity, 
and then build and evaluate such a cavity

\item  Continue to work on the integration and optimization of the rf within
the cooling channel layout

\item  Develop high-power rf windows, couplers and ancillary equipment for
the cavities

\item  Continue to evaluate high-power rf sources and modulators, working
with potential vendors to identify critical R\&D items
\end{itemize}

\begin{figure}[!tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA3.eps}
\end{center}
\caption{Cavity impedance versus length for an ideal pillbox, $\protect\beta $=
0.87.}
\label{APP-RF:fg3}
\end{figure}

\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA5.eps}
\end{center}
\caption[Schematic of multi-beam klystron]{Schematic of multi-beam klystron.\\
 \textbf{http://www.tte.thomson-csf.com}}
\label{APP-RF:fg5}
\end{figure}

\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA6.eps}
\end{center}
\caption{Cathode of Thompson multi-beam klystron.}
\label{APP-RF:fg6}
\end{figure}

%(http://www.tte.thomson-csf.com)
%(http://www.tte.thomson-csf.com)
\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA7.eps}
\end{center}
\caption{Cavity of multi-beam klystron.}
\label{APP-RF:fg7}
\end{figure}

\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA8.eps}
\end{center}
\caption{Klystron efficiency \textit{vs.} beam perveance.}
\label{APP-RF:fg8}
\end{figure}

\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/RF-fgA9.eps}
\end{center}
\caption{1~MW cw HOM-IOT.}
\label{APP-RF:fg9}
\end{figure}

\begin{table}[tbh]
\caption[Comparison between HOM-IOT and klystron]{Comparison between HOM-IOT
(expected results) and klystron, both operated at 1~MW~CW and 700~MHz.}
\label{APP-RF:tb1}
\begin{center}
\begin{tabular}{|lcc|}
\hline
Device & HOM-IOT & Klystron \\ \hline
Effective efficiency (\%) & 73 & 60 \\ 
Assembly volume (ft$^3$) & 30 & 200 \\ 
Assembly weight (lbs) & 1,000 & 5,000 \\ 
DC beam voltage (kV) & 45 & 90 \\ 
Gain (dB) & 25 & 46 \\ \hline
\end{tabular}
\end{center}
\end{table}

\section{Acceleration System}

The most challenging aspect of the acceleration system is the 201-MHz
superconducting rf (SCRF) cavities. The history of SCRF development for LEP,
CEBAF, CESR, KEKB and TTF (TESLA) shows that it takes many years to design,
prototype, and test structures in order to be ready for production. The
lowest frequency at which SCRF cavities have been made for accelerating
relativistic particles is 352 MHz for LEP-II. Therefore, R\&D and
prototyping for a Neutrino Factory at 201.25 MHz has been started now.

At present, SCRF R\&D is in progress to address the following issues:

\begin{itemize}
\item  Achieving 17 MV/m at a $Q$ of $6\times 10^{9}$ in a single-cell
201.25-MHz cavity

\item  Stiffening the 2-cell cavity designs to reduce Lorentz force detuning
and microphonics sensitivity

\item  Exploring pipe cooling, both to reduce liquid-He inventory and to
help stiffen multi-cell structures

\item  Reducing structure cost
\end{itemize}

A collaboration has been set up with CERN to produce a single-cell Nb/Cu
cavity at 201.25 MHz. CERN will provide the copper cavity, coat it with 1--2 
$\mu $m thick niobium film using their standard DC-magnetron-sputtering
technique, and send it to Cornell for testing after high-pressure rinsing
and evacuation. To test the cavity, Cornell is upgrading its test
facilities. Figure \ref{APP-RF:fg10} shows a 3D CAD model of the CERN
cavity inside the test dewar. A test pit 2.5 m diameter by 5 m deep is under
excavation (Fig. \ref{APP-RF:fg11}) to accommodate the test dewar, which has
been ordered. A 201.25-MHz, 2-kW rf test system is under construction. The
clean room and high-pressure rinsing system at Cornell are being upgraded to
accommodate the large cavity. ANSYS calculations have started on the 2-cell
cavities to determine the mechanical resonant modes and frequencies. Not
surprisingly, the resonant frequencies are low. Exploration has started on
stiffening schemes, with and without pipe cooling. Figure \ref{APP-RF:fg12}
compares calculated $Q$ {\it vs.} $E$ curves for pipe cooling and bath cooling
operations. 

At 201.25 MHz, structure costs will be substantial. Multicell cavities are
usually fabricated in parts that have to be machined, cleaned, and
electron-beam welded. This is an expensive, labor-intensive process. We are
collaborating with INFN-Legnaro in Italy to spin monolithic copper cells out
of a single tube. Legnaro has experience at 1300 MHz. As a first step, they
will spin a single-cell 500~MHz cavity. In a future stage, the procedure
will be extended to 201.25~MHz and multi-cell cavities.

\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/rf_fg31.eps}
\end{center}
\caption{Vertical dewar test.}
\label{APP-RF:fg10}
\end{figure}

\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/rf_fg32.eps}
\end{center}
\caption[Test pit (2.5~m diameter and 5~m deep) for 200~MHz cavities]{200~MHz test pit
(2.5~m diameter and 5~m deep) under construction at Cornell. The other pits
are for testing existing cavities.}
\label{APP-RF:fg11}
\end{figure}
\begin{figure}[tbp]
\begin{center}
\includegraphics[width=3in]{../template/report/ps-and-eps/rf_fg33.eps}
\end{center}
\caption[Comparison of pipe-with bath-cooling for a 200 MHz cell]{Comparison
of pipe with bath-cooling at 2.5 K for a 200 MHz single-cell cavity. The
He-carrying pipe diameter is 10 mm; spacing between pipes is 70 mm.}
\label{APP-RF:fg12}
\end{figure}

One long-term goal of the R\&D is to design, construct, and high-power test
a cryomodule with the first single-cell 201.25-MHz cavity, equipped with
couplers and tuners. To prepare this test, continuing R\&D, design, and
prototyping are necessary in the following areas:

\begin{itemize}
\item  high-power input coupler

\item  higher-order-mode coupler

\item  mechanical/thermal tuner

\item  piezoelectric/magnetostrictive tuner

\item  cryomodule

\item  system integration

\item  high-power testing
\end{itemize}

Future R\&D on structure stiffening, feed-forward, and active tuning to
compensate Lorentz force detuning and microphonics could lower the required
peak power by reducing the detuning tolerance. For example, if the detuning
tolerance can be lowered to 20 Hz, the input power drops to 450 kW per cell
and the optimum $Q_{L}$ increases to $1.5\times 10^{6}.$ Adopting a 4 ms
fill time would then decrease the input power requirement to 350 kW per cell
at the best $Q_{L}$ of $1.5\times 10^{6}$---a level already reached at KEKB.

The acceleration system arc design, while reasonably straightforward,
requires a number of nonstandard components. Design concepts for the
injection chicane and the arc magnets are needed. Depending on their
complexity, prototypes might be needed for some of these.

\section{Storage Ring}

The arc magnet concept proposed here is novel, and a prototype device is
certainly called for. In addition to evaluating the coil fabrication
aspects, measurements of field quality suitable for the tracking studies
must be performed. Thereafter, the tracking must be carried out to ensure a
design with acceptable dynamic aperture for injection and storage.

Optics designs to reduce or eliminate the contributions to the detector from
the ends of the straight section, where the Twiss parameters are not
suitable in terms of the beam angular divergence, must be done. It should be
possible to ``hide'' the matching regions from the detector with suitable
horizontal or vertical bends, but this must be verified with an actual
lattice design. 

Finally, the cost-benefit tradeoffs between the present compact design and a
conventional ring with a liner to protect the magnets from beam decay
products must be quantified.

\begin{thebibliography}{99}
\bibitem{RandD:ref1}http://www.cap.bnl.gov/mumu/RandD/RandD\_R3.pdf

%\bibitem{RandD:ref2} Need this reference

\bibitem{Diesburg-FMC}  D. Diesberg, Lithium Division, FMC Corp, private
communication.

\bibitem{Brush-Wellman}  J. Crim, Brush Wellman Corp., private
communication. %\begin{thebibliography}{9}

\bibitem{mike:ref1}  A. Beunas, G. Faillon, THOMSON TTE, France, S. Choroba,
A. Gamp, DESY, Germany {\sl A High Efficiency Long Pulse Multi Beam Klystron
For The Tesla Linear Collider}, Submitted to PAC 2001.

\bibitem{mike:ref2}  H.P. Bohlen {\sl Advanced high-power microwave vacuum
electron device development}, Proc. 1999 Particle Accelerator Conference,
pp.445-449. %\end{thebibliography}
\end{thebibliography}

%\end{document}