here are the minutes of Tuesday's meeting. If you find that my personal
bias led me to forget or misrepresent anything, please send an email
to the distribution list to put me straight.
I think we had a rather productive meeting (if that isn't an oxymoron).
While we haven't exactly pinned down all the design details, we have
made some progress toward what I would call a "reference design". This
will serve us as a starting point; any changes should come from intense
scrutiny of this reference and hopefully can be incremental.
Starting from the front to the back, I describe the general outline
of the target as the beam (and scattered particles) see it:
*****
The first item we talked about is the beam line entering the target.
We agreed it should have a gate valve at the upstream end, and attach
directly to the target vacuum chamber downstream. The vacuum chamber
entry window will separate the beam line vacuum from the target vacuum;
meanwhile, this piece of beamline can be pumped out and then connected
to the rest of the accelerator vacuum by opening the gate valve.
ACTION NEEDED: We need to design the entrance window. It will be =
straightforward, maybe 2 mil of Al or 1 mil Titanium, 4 cm diameter.
After entering the target vacuum chamber, the beam will go through a
tube that protrudes into the LHe target cryostat. The downstream end
of this tube will have to have a superfluid-tight window of about 4 cm
diameter. One possibility would be 1 mil of Ti, welded to the rim
or machined directly out of a rim piece of the tube. Up to this point,
the beam has then seen about 0.1% of a R.L.
ACTION NEEDED: Design of the end piece of this tube with window.
The beam then goes through the LHe of the cryostat and the target cell
itself. Our reference design for a target cell is the following:
a 1 cm long, 1.5 cm diameter cylinder made of Kevlar, Torlon
or similar material, with (thin) Al entrance and exit windows,
with an NMR coil embedded in the target material inside, but in a
"Helmholtz"-type arrangement which minimizes the coil material within
the central 1 cm diameter of the cylinder (the coils will be shaped
in a "drooping" fashion around the central 1cm dia. cylinder). We can
then choose to either raster over the whole volume (gives more accurate
polarization measurement) or over the central 1 cm only (better dilution
factor). =
We decided to have 4 target cells: one ^15NH_3, one ^14ND_3, one con-
taining the "background" material (see below) and one empty one (which
can be used to "park" the beam, or for additional background studies.
Possibly one could even hook it up to NMR as well and put a second
NH3 target in there). We plan to foresee the possibility to anneal
the targets during the run (may be needed once or several times per
week).
ACTION NEEDED: Both UVa and Genova will try to build prototype cells
following these general outlines. We need to study whether we can
really get the coil (mostly) out of the way, whether the NMR signal
(TE) will be large enough, what packing fraction we can get with these
rather small cells, etc. It will be hard to accomodate longer target
cells, since that would give us too big a beam spot downstream. The
present design is already cutting it close.
The 4 target cells will be spaced at 2.5 cm distance along the target
stick, so that an overall travel of 7.5 cm (+ additional room for
alignment) is needed - well within the maximum of 10 cm. We MIGHT
consider adding a "ring" of 1.5 cm right below the lowest cell to
have another spot for the beam to "park", which would increase the
travel distance to 9 cm + alignment.
The trumpet for the =B5wave will come in from behind, snake around and
blast the cell from the side (about 90 degrees). Marco and Don made
a visit to a company that claims they can fabricate the corresponding
wave guide; it is probably impossible to move the trumpet into the
shadow region at 60 or 120 degrees. However, because the target volume
is rather small, the trumpet can be small, too, and won't obstruct
too much of the sideways openings.
The target material for the "background" target is still open to dis-
cussion. Ideally, it would be N2. However, there may be difficulties
associated with this (it obviously freezes below LN2 temperature - LNe
would work), and we might not know how much we got into the cell. =
Alternatives range from conservative (Carbon, CO2) through
exciting (N2O) to exotic (N2H2 =3D hydrozine?, N3H =3D hydrolysic acid ??=
).
Mike Seely (and perhaps others?) will look into the latter materials,
to determine freezing point, density, radiation properties (if known),
and possible hazards (explosion, poisoning, uncontrollable laughter).
It would be better if we wouldn't have to introduce NEW nuclear species
(like O, C) into our background target.
In any case, I estimate that the beam will be going through an extra
0.5 - 1cm of LHe in front and behind the target cell, since the
windows (especially the outgoing one) will be bulging, and we need
enough clearance for the vertical motion. All in all, I expect the
target plus LHe plus entry windows to add up to about 1.8% of a
radiation length. We need to try and stay below that for the rest
of the material that the beam and the outgoing particles have to
go through. Specifically, we have 3 large windows (exit window of
cryostat, heat shield, and vacuum exit window) in forward direction
(which have to accomodate up to +/- 50 degrees in theta without
shadowing anything beyond 7 degrees) and the same in the 6 sideways
ports. =
Coming to the latter first: The 6 sideway ports on the cryostat
chamber will simply be machined-down sections of the (stainless steel)
walls. Hopefully they will be thin enough (<5 mils =3D 1% R.L.)
ACTION NEEDED: Marco (or Don or Mike?) will contact Oxford to determine
what the present dimensions of these sideway ports are, and how thin
we can have them shave them.
The front window will have to allow for the beam AND the scattered
particles up to 50 degrees to go through unshadowed. This precludes
the use of any "spokes" or "rings" on this window. However, we can
design a lid (2 possibilities were shown at the meeting) which brings
this window rather close to the target cell exit (maybe 0.5 cm), which
is desirable anyway. In that case, a 5-6cm diameter would be sufficient,
which could be accomodated by another 1-2 mil Ti foil.
ACTION NEEDED: We need to design and draw the lid and decide how
the foil will be attached without any bolts/rim getting in the way
of vertical target motion.
The first really big window will be in the heat shield. Oxford may
have a design already, but most likely it won't fit our bill. Again,
there are side windows for the 6 side ports, plus a huge (about 67 cm
diameter) forward window. The problem is that this window needs to =
have good enough heat conductance to cool it to near LN2 temperatures
right in the middle, where the beam goes through. Obviously, we could
make some composite design, with a large, somewhat thicker (1 mm?)
plate having a small (4 cm dia) hole in the center, which in turn is
covered by a thinner foil for the beam to go through. Fortunately this
is not a vacuum/pressure window, so few mechanical constraints exist.
It was also suggested that one could add "fingers" to the window
that are much thicker and lie in the shadow of CLAS' acceptance;
these could aid with the heat transport away from the center. However,
we can NOT have a central ring, since it would either clip the beam
or shadow some of the 7 degree scattered particles. Another idea
was to reduce the need for a truly thick heat conductor by having about
10 layers of superinsulation (that adds about 1mil) between this
heat shield and the outer chamber surface.
ACTION NEEDED: Obviously, here we need major design/prototyping work.
We also need to tap the experience of seasoned cryogenic target builders
(Hall C/A ?) and Oxford to assess what combination of thin foil,
fingers and superinsulation works best and lets us get away with the
least amount of material in the way of outgoing particles. Maybe we
will need to do tests with different configurations, using the real
target and reducing the heat shield thickness until the boil-off rate
becomes prohibitive. Don and Mike can help with info (and contact
Oxford to see what they've designed so far and what they think might
work), but clearly we need more help on this one.
Finally, we need the large vacuum exit window on the downstream side
of the can. (Again, we also need 6 side windows; however, they will
be relatively small and shouldn't pose huge design problems. That doesn't=
mean we DON'T need to start working on their design now!). Here, we
need to cover about 80 cm diameter to withstand 1 atm +. Again, the
best solution is probably a composite, involving "spokes" (and here we
could have a small "ring" right outside of 4 cm diameter without
shadowing 7 degrees), a thicker large window (made out of a low-Z,
lightweight material - Kevlar, Mylar, you name it) with a hole covered
by a smaller (2-3 mil) Al foil (this foil can and SHOULD cover the =
WHOLE outside of the window for electric and light shielding.
ACTION NEEDED: Again, major design, test and fabrication work. Needs
as much input from experts as we can get.
*****
Miscellaneous other items:
1) You'll find some action items above that don't have a specific
name attached to them. Of course, that should encourage you to take
over the task in question. However, we seriously need some engineering/
drafting manpower to design and draw all the parts we need. Unfortunately=
,
it didn't become clear to me at our meeting who this could be, or where
the manpower will come from (note that Walter Tuzel is no longer
available). Certainly, it will HAVE to be someone stationed at JLab
working closely with Mike Seely.
2) Volker reported that a straightforward beam transport calculation
showed that the solenoidal field from the target can wreak a lot of
havoc on the electron beam optics, especially at the lowest energy
(1.6 GeV). We need more detailed studies and consider possible
remedies (i.e., altering the incoming beam optics and/or the rastering
scheme) to at least partially compensate for this.
3) Mike reminded us that alignment will be a recurring problem unless
something is done to the insertion cart rails that would allow us to
move the target in in one smooth move. Right now, the rails are crooked,
which requires several LATERAL adjustments to be made during insertion
(to avoid bumping into the R1 drift chambers). We should discuss with
the Hall engineers (O'Meara?) whether and how the rails could be
shimmed/readjusted to make this possible. Our only alternative is to
have an alignment crew come in EVERY time we move the target in or out.
(We MIGHT be able to avoid that if we put reference marks close to the
once-surveyed "in" position; at least, this might speed the process up).
This reminds me that we haven't really understood how Oxford is proposing=
to reference the position of the magnet relative to the outside of
the can, so we don't have to insert the target "open", but rather can
use survey marks (tooling balls etc.) on the outside for final alignment.=
4) Now that we have a "reference" design, we can implement it in our
favorite CLAS Monte Carlo and see what all that material does to
our resolution, acceptance and reconstruction. We also still need
to look into optimal target position in z (back from the nominal
target point), CLAS field strength and polarity, etc. We should have
another meeting "soon" - maybe in late October - and hopefully some
results/answers to these questions.
OK, this is probably at least MY personal record for a lengthy email
message. 'nuff already! You can always erase it - our distribution
list keeps an archive, so you can go back there to read it again later...=
Greetings -
-- =
- Sebastian