In building any detector for and experiment, one usually finds that all aspects of the detector are constrained and in many cases over constrained. Thus the first, and oft times the hardest tasks, of the designer are to determine all of the relevant constraints, assess the importance of each of those constraints, and make the appropriate compromises where necessary. For the E896 tracking detector, most of the outer dimensions are only constrained by the 48D48 Analyzing Magnet which was chosen because it provides a sufficeintly large fiducial volume in which to observe H0 decays and identify the resulting daughter particles. More specifically, the tracking detector in total (electronics, cabling, etc) is constrained to be not more than 45cm in height, 120cm in width, and 120cm in depth. By contrast, the inner or "working" area of the tracking detector was determined by a of combination of hardware constraints imposed by the Sweeping Magnet and background simulations obtained using GEANT. The inner height of 20cm is a compromise between opening necessary to avoid photon conversions in the outer shell of the tracking detector and the finite space required for structural integrity of that shell and the necessary on board electronics. Note, the 15cm height of the sweeping magnet is itself the result of the compromise between achieving the highest sweeping field possible while still allowing room for silicon vertex detector, SVT. The 67.5cm inner width of the tracking detector was determined by the afore mentioned background Monte Carlo. The results indicate that from the zero degree beamline to 25cm beam left should be active to detect H0, 20cm to beam left of this active region must remain empty to avoid secondary collisions from pi-minus particle, while 22.5 cm to beam right of zero degree beamline must remain empty to avoid secondary interactions from uninteracted beam and projectile fragments.
A 2mm double track resolution is required from the tracking detector transverse to the direction of the beam to resolve the Sigma-minus and proton channel produced in the one of the more interesting of the H0 decay channels. The required granularity along the beam direction (Z) was determined to be ~1cm; a compromise between need for multiple measurements along Sigma-minus track, the need to maximize the depth of the detector to increase the number of H0 decays observed, and the availability of electronics with which to outfit instrument.
Given these restrictions, it was determined that a distributed drift chamber (DDC) was probably the optimal detector for this experiment. It was calculated that simple Multiwire Chamber configuration would lack the spatial resolution (approximately a factor of 10 less than DDC) necessary to reconstruct invariant mass of Sigma-minus Proton channel and thereby distinguish it from the K-minus Proton background produced in neutron neutron interactions. A TPC approach was not chosen due to insufficient double track resolution and because typical recovery rates are much slower than a DDC (~100). A Drift Tube configuration was rejected because its grammage is larger than DDC (13:1) as well as the fact that its assembly would be significantly more difficult than a DDC.
Once the overall size, shape, and type of tracking detector had been decided, investigation into specific construction materials and the actual internal configuration was begun. It was decide that the cathode planes would be composed of wires not foils. Use of aluminized mylar foils would present a greater possibility of electric field problems (e.g. break down, polymerization of wires) at edges of foils since we would not able to extend the foils into pi-minus or beam regions. The use of foils would also place more stringent requirements on the flatness and "parallel"-ness on the NEMA G10 cathode mounting frames. finally, use of foils might introduce gas flow problems or at the very least require longer purging time. Thus the cathode planes will be composed of 75 micron electro-polished 316 stainless steel wires. Literature shows that stainless steel wires retain their tension unlike Aluminium. Aluminium also oxidizes which tends to insulate the wire leading to the "Malter" effect. Similarly Be:Cu (mostly Cu) is more susceptible to electron emission than iron. By contrast the all but the terminating anode/sense wires were chosen to be 20 micron gold plated tungsten. The tensil strength of tungsten allows for the smallest diameter wire, hence greater electric fields with lower voltages. The gold plating provides smoothness which makes for more uniform field over the length of the wire. It also allows the wires to be soldered. The terminating anode wire on either end of the sense planes is 75 micron electropolished 316 stainless, like the cathode wires. These outside anode wires must be larger diameter than central anode wires for proper field termination.
Although the portion of the chamber where beam and projectile fragments pass will not be active it will still contain chamber gas. Thus it is preferable to use a Helium-based gas to reduce the amount of delta rays and secondary interactions produce in the gas by the uninteracted beam and projectile fragments. We have decided to go with He:C2H6 (50:50), bubbled though methylal. It is expected that the effect of the magnetic field on the drift velociy of the "cooler" helium-based gases will be less than that experienced in the "hotter" Argon-based gases. Similarly, the effect of the magnetic field on pulse height should also be reduced. The introduction of 1-5% Methylal is known to significantly reduce or negate polymerization of wires (aging) in chamber.
The internal cell configuration was optimized by performing extensive GARFIELD simulations. The configurations which achieved the best spatial resolution were the ones in which the ratio of the cell width to depth was maximized. Since the maximum width is set to 2mm by the double track resolution, the only option in maximizing the width:depth ratio was to minimize the cell depth. The cell depth was determined to be 3mm, which is minimum thickness with sufficient strength to support and maintain tension for 125 Stainless Steel wires (~5kg). This cell layout provides each sense plane with two dedicated cathode planes such that no two sense planes need share a common cathode. This should remove any electric field problems incurred with shared cathode planes because now the orientation of wires in companion cathode planes are matched to the sense wire orientation. This 4mm*6mm cell showed a "bleed" over of drift electrons to neighboring wires for only about 20% of the outer most tracks. Further GARFIELD simulations were run producing plots of lines of equipotential in the presence of no magnetic field, drift lines in the presence of a 1.8T field, and drift distance versus time.
The wire orientation is currently envisioned as 0 degree (X), +15 degrees (T), and -15 degrees (U). However, this is still an open parameter at this point and is currently being optimized using tracking simulations. Similarly the plane sequencing, now XTUX'T'U', is also considered open and is currently being optimized. Neither of these should is seen as an impediment to the completion of a chamber design at this juncture.
