Constructing individual calorimeter towers with the
dimensions 10 cm by 10 cm by 100 cm would be most convenient. A cross-section
of such a tower, assuming 1 mm diameter fibers and the
lead-to-scintillator volume ratio of about 4:1, is shown in
Fig. 
. Each tower pictured contains 2500 fibers.
We investigated casting each tower, however this method proved to be too difficult. The 1 mm diameter holes, which were cast in the tower with stainless steel rods, inevitably varied from their ideal positions as the rods ``clumped" together due to their small diameter, close proximity, and meter-long length. The ``clumping" would cause certain regions of the tower to be over-sampled with fiber, with other regions lacking the proper density of scintillator. This nonuniformity would detract from the calorimeter's resolution. Another technique casts scintillating fiber directly in a low melting point eutectic alloy. However, the alloy contains a large percentage of cadmium, which has a large neutron cross section, and is thus not acceptable for hadronic calorimetry.
A more widely-used method is that of constructing plates with
half-circular grooves, and then bonding the plates together. Grooves can be
extruded, rolled, or machined into the plate. An example of
such a plate is shown in Fig. . It would be easiest to
set the fibers into their grooves, spread a layer of epoxy over the lead and
fibers, and then sandwich this with another plate of lead. This method, though
simple, is irreversible and can potentially exacerbate the radiation damage in
the fibers.
The SPACAL collaboration at CERN is developing an alternative technique where extruded, grooved plates are tinned with a thin layer of solder, and then the plates are stacked and heated until the solder flows. The plates are thus bonded together but the holes remain empty so that fibers can be inserted after the tower is constructed. This has the advantages that the fibers can be removed from the module, and there is an air layer around the fiber which makes the fiber more resilient to radiation damage [73]. Unfortunately, this technique increases the cost of the extruded plates by a factor of five [74].
E864 is also developing a technique to construct towers out of grooved plates. As in the SPACAL method, the goal is to construct a tower which contains empty holes for the fibers, however, the adhesive used in this method is a 3M heat-curing epoxy [75]. This epoxy, which comes dissolved in a solvent, can be painted on material and left to dry to a tack-free surface. After painting a thin layer of epoxy on a plate of lead, grooves would be machined in the lead, thereby removing both the lead and adhesive from the groove. The machined plates would then be stacked and placed in an oven to cure. As in the SPACAL method of bonding, the fibers can be removed from the module, and there is also an air layer around the fiber which helps the fiber resist radiation damage. Applying the heat curing epoxy to the lead plates should only add about 10% to the cost of the plate.
Figure: Cross-section of a single calorimeter tower using 1 mm diameter
fibers and a lead-to-scintillator volume ratio of about 4:1. (Dimensions are in cm.)
Figure: Cross-section of a grooved plate. Fifty stacked plates will make a
single tower. (Dimensions are in cm.)